SURP 2024 PROJECTS
Below you will find a list of available 2024 CENG Summer Undergraduate research projects. Click on the plus sign of any project title to learn more about the project. You can apply for up to 3 projects and need not be in the same department as the advisor(s).
If you have any questions, reach out to Dr. Bridget Benson, Associate Dean of Scholarship and Assessment at bbenson@calpoly.edu
Faculty: Dr. Ayaan Kazerouni
Department: CSSE
Email: ayaank@calpoly.edu
Co-Advisor: Dr. John Clements, CSSE
Accepted Projects Modes:
Hybrid
Software testing is a core competency in undergraduate computer science and software engineering programs. As such, students are encouraged to use automated software tests to assess the correctness of their own programs, and in turn are encouraged to use test adequacy criteria to assess the quality of their automated tests. Two prominent examples of adequacy criteria are code coverage, which is commonly used, computationally inexpensive, and easy to reason about, and mutation analysis, which is often more sensitive, but can be expensive to compute and challenging to reason about as a form of feedback. Mutation analysis works by systematically creating faulty versions of the program-under-test (“”mutants””), and evaluates the quality of the automated tests by running them against the mutants. If any test fails, it means the mutant was detected. Mutants that “”survive”” represent gaps in the automated tests that ought be to be addressed. Numerous schemes have been proposed to reduce the computational and cognitive cost of mutation-based testing feedback. The PI has engaged in existing scholarship related to mutation analysis and its efficacy as a tool for teaching software testing. In this SURP project, we aim to continue this work to answer the following research question: When a test suite satisfies a weaker criterion (like code coverage), what number and type of mutants tend to go undetected? We note that mutation analysis subsumes code coverage. This means that, once a test suite satisfies code coverage, the mutants that are leftover may represent the “”gap”” between the two test adequacy criteria. Understanding which types of mutants tend to “”survive”” the weaker (but cheaper and simpler) criterion would allow us to only evaluate test suites using those types of mutants. This could leading to significant savings in terms of computational cost and developer testing effort. We plan to investigate this question empirically using data from two sources: (1) Programs and test suites created by students here at Cal Poly, in courses like CSC 305 or CSC 430, and (2) Open-source software and test suites publicly available on GitHub. The proposal asks for two students to work on this project, so that similar analyses can be carried out on multiple corpuses of programs in different programming languages. This will lend some generality to our findings.
Faculty: Dr. Chokri Sendi
Department: AERO
Email: csendi@calpoly.edu
Accepted Projects Modes: Hybrid
This project investigates the attitude stability of a flapping-wing UAV made of a rigid platform and flexible wings during low-speed retargeting maneuver. In addition to the attenuation of external disturbances and model uncertainties, the designed fuzzy controller goal is to suppress the vibration caused by the two-dimensional flapping dynamics of the wings, and to stabilize the attitude of the UAV subject actuators saturation. The fuzzy controller-observer is formulated as a set of linear matrix inequalities and utilizes the parallel distributed compensator to obtain the feedback control and the observer gains. A prototype of the UAV will be tested to validate the dynamic model, and the fuzzy controller. The outcome of this project will be published in a reputable journal, and it will be used as a basis to seek funding from government agency such as NSF and NASA.
Faculty: Dr. Eric Mehiel
Department: AERO
Email: emehiel@calpoly.edu
Accepted Projects Modes: Any
With the significant growth of the commercial space industry comes questions of the ethical use of space as a platform for commercial business. In addition, while space has traditionally been the domain of sovereign nations, often guided by international treaties, the current growth of the commercial space industry is highly intertwined with national budgets, policy, and military spending; particularly in the United States. Even further into space, there are questions regarding the accumulation of solar resources contained on the planets, the Earth Moon, and asteroids. For example, “who owns the moon?” What laws, treaties, or agreements govern the use, gathering, and sale of these resources? What does our colonial/capitalistic past and present tell us about the
ethics of these choices we are confronted with? This project will start with basic research into the history of space utilization (including relevant policies, treaties and laws) and work toward developing/applying an ethical framework to the use of space resources.
Faculty: Dr. Eric Mehiel
Department: AERO
Email: emehiel@calpoly.edu
Accepted Projects Modes: Any
With HSF User Interface development in progress, the next step to support our goal of opensource community development for HSF is to start developing a library of system models. The goal is to create a variety of parameterized models of space systems that can be imported and combined to support rapid spacecraft/system design and test with HSF. Currently, only basic spacecraft components have been modeled so the potential list of models that can be developed is large. This project will offer students an opportunity to design and implement a wide range of spacecraft models which will involve applying what they have learned about spacecraft subsystems in support of Model Based Systems Engineering.
Faculty: Dr. Eric Mehiel
Department: AERO
Email: emehiel@calpoly.edu
Accepted Projects Modes: In-Person
There are three typical methods used to control the attitude (or orientation) of spacecraft in orbit. They are momentum exchange devices (reaction wheels and control moment gyroscopes), magnetic torquers, and thrusters (both cold and hot gas). The PolySim research group is currently developing a spacecraft attitude dynamics simulator utilizing reaction wheels. The simulator floats on a spherical air bearing and is balanced to accurately simulate the rotational motion of a spacecraft in orbit (not including the orbital motion). As a continuation of that project, we have started development of a spacecraft attitude simulator using cold gas thrusters to control the altitude of the simulator. The goal of this project is to design, develop, manufacture, and test a single axis attitude control simulator using thrusters. The thrusters will be fed with compressed air and are ‘cold gas’. The project will require completion of several tasks. First, the mechanical and the electrical systems must be developed and designed. This includes the microcontroller and the solenoid system used to fire the thrusters. Next, the student will design a feedback control law using pulse-width modulation for the ‘bang-bang’ thruster implementations, including simulations in Matlab and Simulink. Next, the development of the microcontroller used to implement the control law and all code must be completed. Finally, the full system will be tested with hardware results compared to simulations developed in Matlab and Simulink.
Faculty: Dr. Stephen Kwok Choon
Department: AERO
Email: skwokcho@calpoly.edu
This research project’s goal is to explore CubeSat actuation mechanisms and end-effectors that can be utilized for orbital debris collection. Orbital debris is projected to increase, due to increased activities such as the insertion of satellite mega-constellations and the risk of impact that such systems pose to increasing the space debris propagation. This Summer Undergraduate Research Project is proposed for students to study and determine how CubeSats using actuation mechanisms and end-effectors can be utilized in achieve active debris removal. It is expected for the student/s to explore current mechanisms that exist as well as propose an approach. This work shall involve rapid prototyping that involves creating a model through additive manufacturing (3D printing), and actuation through mechatronics (servos, DC motors, actuation mechanisms). I am projecting for the student/s to have a working prototype of their CubeSat Mechanism/s at the completion of this Summer Undergraduate Research Project.
Faculty: Dr. Stephen Kwok Choon
Department: AERO
Email: skwokcho@calpoly.edu
Accepted Projects Modes: In-Person
Application Link
This project aims to explore the sensor and actuation mechanisms necessary to enable docking recharge stations for autonomous rover vehicles on the lunar surface. The concept being proposed would be solar-powered recharge stations that would extend the operational range for autonomous rovers to function. Recharge stations are currently being explored to operate on the Lunar Surface [1–3]. Recharge hubs would allow rovers to operate within a safe zone, thereby returning to the recharge hub as needed[4–6]. Recharge Hubs could potentially reduce the complexity and cost of rover vehicles because they would not need to carry as much on board (batteries, solar cells, etc). Recharge hubs also could be potentially chained allowing rovers to operate in a linear path from one site to another[7–10]. The applications of using recharge hubs are linked with resource gathering and collection. This Summer Undergraduate Research Project aims to advance knowledge in how an integrated network of Rover vehicles could operate on the Lunar Surface performing tasks that would otherwise be too dangerous, tedious, or not feasible for humans to perform.
Faculty: Dr. Britta Berg-Johansen
Department: BMED
Email: bbergjoh@calpoly.edu
Accepted Projects Modes: In-Person
The main goal of this project is to bring a “mobile biomechanics lab” using smartphones and smartwatches into local schools to conduct biomechanical research studies and provide hands-on interactive biomechanics lessons for children from groups traditionally underrepresented in research. This project has openings for two student roles:
Role 1 will focus on outreach and establishing collaborations with SLO county middle and elementary schools, which – like many in the state of California – serve a large population of Hispanic families. Due to historical and institutional factors, these families are likely to include a substantial population of socioeconomically disadvantaged children who are at higher risk of experiencing high rates of overweight
and obesity. The research student in this role will also lead the development of educational hands-on biomechanics lessons for the children. This role is open to all engineering majors, with Spanish speakers preferred.
Role 2 will focus on (1) further developing our lab’s innovative methodologies to measure balance and gait metrics using smartphones and smartwatches and (2) conducting a literature review on inclusive research practices and product design in order to help revise our participant research study practices and surveys with a diversity and inclusivity lens. The methodologies and practices developed will be used in the educational biomechanics lessons described above and will also be used in our lab’s current and future research studies, such as investigating the effects of body mass index (BMI) on balance and gait biomechanics in children. The student in this role will also help develop a workshop with Dr. Berg-Johansen on inclusive wearable device design for students and faculty in the next academic year. This role is open to BMED, ME, and CPE majors, with preference given to those with experience using MATLAB for data analysis/processing.
Students should indicate in their application which ole(s) they are applying for and whether they are interested in continuing work on the project during next the Academic Year through independent study courses
Faculty: Dr. Michael Whitt
Department: BMED
Email: mdwhitt@calpoly.edu
Accepted Projects Modes: In-Person
Endothelial dysfunction is the most significant predictor of a major adverse cardiovascular event (MACE). It is also associated with other diseases including COVID-19 and diabetes. Greater understanding of this metric in communities where diseases associated with endothelial dysfunction are prevalent could lead to early detection and better patient outcomes. The research work performed will increase student knowledge in areas including development of Institutional Review Board protocols, how to perform research in a clinical environment, and potential interactions with groups previously in communication with Whitt. One such group is the FAITH! Study in conjunction with the Mayo Clinic where African- American churches play a key role in educating about cardiovascular disease in the African American community. Another potential group is the long COVID group at UCLA Health. There are other potential groups and diseases that will be explored with the students in line with their interests and passions. Any knowledge gained about endothelial dysfunction associated with any diseases studied could positively impact future patient outcomes. Other cardiovascular metrics will be explored as well as part of this study of noninvasive metrics.
Faculty: Dr. Shreeshan Jena
Department: BMED
Email: shjena@calpoly.edu
Accepted Projects Modes: Hybrid
The combination of outsole, midsole and insole layers of new shoes provide proper cushioning to the foot during the daily activities of life. Subsequently, the wear-based degeneration or damage to one or more of these sole layers can reduce the cushioning effect to the plantar surface of the foot. This reduction of cushioning effect can impact the distribution of stress concentration regions along the foot plantar surfaces and lead to foot pathologies. The present proposed research intends to study these effects of shoe-wear on the plantar surface of the human foot, using FEA (ABAQUS) and experimental methods. Commonly, the worn-out zones of shoes are observed as flattened shoe soles around the forefoot region, heel, medial or lateral zones. The proposed research intends to simulate the effect of gait-based loading on a human foot geometry, when applied through the multiple layers of worn-out soles, compared to that of a new shoe sole geometry. The gait-based loading on the geometry shall be obtained from the data captured from a force platform.
Faculty: Dr. Derek Manheim
Department: CEENVE
Email: dmanheim@calpoly.edu
Accepted Projects Modes: In-Person
The rapid increase in global plastics consumption along with dwindling recycling rates worldwide have caused great concern over the fate and persistence of microplastics in the environment, including potential contamination of food supplies and water resources. Microplastics are defined as any plastic material with size dimensions less than 5 mm, including nanomaterials, that may be discharged directly to the environment through waste disposal and/or generated in the environment from macro-plastic pollution. Dietary ingestion of microplastics and subsequent chemical exposure induced from these materials is of growing concern, particularly when waste management activities, such as land application of treated wastewater sludges or digestate, directly recirculate and remobilize concentrated amounts of these materials back into the environment. Therefore, monitoring the fate and behavior of microplastics in engineered waste treatment systems, such as dry anaerobic digesters, is increasingly important to devise strategies to optimize their removal and management prior to environmental release. This experimental research aims to: i) quantify the persistence of microplastics within dry anaerobic digesters and ii) evaluate their influence on the treatment performance and renewable energy generation potential at scale. In collaboration with the SLO Kompogas Digester facility, samples will be acquired, and bench scale experiments will be conducted to track the quantity, types, and quality of microplastics persisting during anaerobic dry digestion. The behavior of these microplastics, in terms of the treatment performance (i.e., stabilization of organics) and biogas generation, will be further characterized. This experimental data will provide insight into what operational parameters can be modified to simultaneously promote microplastic sequestration/removal, prevent mobilization/reactivity, and ensure acceptable treatment/operating performance. Ultimately, this data will inform the optimization of future operational practices at the full-scale Kompogas digestion facility to best manage, treat, and/or sequester microplastics from the digestate and composted solids streams prior to discharge into the environment.
Faculty:Dr. Derek Manheim
Department: CEENVE
Email: dmanheim@calpoly.edu
Accepted Projects Modes: In-Person
The increasing frequency and intensity of catastrophic natural disasters disproportionally impacts the livelihoods and wellbeing of low income and minority communities. The overwhelming generation of post-disaster materials following these events places immediate stress on waste management authorities to ensure safe, efficient, and sustainable material removal, storage, and processing from affected communities. Social equity concerns during post-disaster materials management operations are increasingly prioritized for equitable distribution of resources to minimize adverse environmental and human health impacts resulting from the materials. In previous research conducted by the PI, a quantitative decision support framework integrating social equity analysis was developed to assist local, state, and federal waste management authorities when planning for and responding to natural disasters. This framework directly integrated social justice metrics on the census tract level (e.g., US EPA’s Environmental Justice Screening Tool) to quantify the demographics of affected communities. However, this approach is over-generalized and does not consider the underlying community and disaster specific factors that elicit impacts to social wellbeing, particularly in disadvantaged communities. Ultimately, social wellbeing and associated behavioral responses in these communities directly impact the realization of beneficial waste material reuse and recycling opportunities. Therefore, this proposed research will develop and integrate a local scale mental health severity index (MHSI) calculation module to evaluate social and behavioral impacts to disadvantaged communities throughout disaster response and recovery phases. A Markov state transition modeling framework will be developed, calibrated, and applied to simulate mental health wellness states of affected community members. The MHSI calculation module will then be integrated into the existing decision support framework to offer improved insight into the social wellbeing of post-disaster disadvantaged communities. Application of the improved decision support framework will inform post-disaster waste management authorities of specific locations to target response/recovery efforts and time periods over which activities must be completed to mitigate adverse impacts on the livelihoods of marginalized community members.
Faculty: Dr. Long Wang
Department: CEENVE
Email: lwang38@calpoly.edu
Accepted Projects Modes: In-Person
This interdisciplinary research, in collaboration with Sony, aims to improve the fit and comfort of socket prosthetics for amputees by utilizing sensing technology and data analytical techniques. Many amputees face issues with prosthetic fit, which can lead to discomfort, pain, and even tissue damage. Our goal is to address these problems by developing a low-cost, universal, and wearable sensing system that can continuously monitor the pressures at the residual limb and prosthetic socket interface. This product will provide feedback to the user, allowing for real-time adjustments, ensuring a comfortable fit, and avoiding injury. Our research group has extensive experience in wearable sensors and machine learning/artificial intelligence. Through our current collaboration with Sony, we plan to integrate Sony’s electronic boards with wearable sensing systems and investigate the performance of different data analytical algorithms. The students will have the opportunity to closely work with Dr. Wang’s interdisciplinary research group and Sony’s technical team.
Faculty: Dr. Giovanni De Francesco
Department: CEENVE
Email: gidefran@calpoly.edu
Accepted Projects Modes: In-Person
This research project has two major objectives. The first consists of enabling Cal Poly earthquake simulator at the Parson’s Laboratory, currently not fully operational and historically rarely used for large-scale testing of structures, to make a significant impact in advancing education and research excellence. The second consists
of providing experimental evidence of the economic impact and losses due to earthquake events on residential wood houses with typical structural and non-structural details used in California. The research project will involve a series of dynamic earthquake simulations on a full-scale two-story wood house incorporating typical
structural and non-structural components of California residences. The research team currently consists of three Cal Poly graduate students and two academic collaborators, Prof. A. Palermo, University of California San Diego Structural Engineering Department, and Prof. M. Li, University of British Columbia, Faculty of Forestry. Students will participate in hands-on activities such as construction, instrumentation, and preliminary testing of a two-story residential wood house. Students will also get the opportunity to participate in a series of hands-on activities that will aim to enhance dynamic experimental testing capability at the Cal Poly Parson’s laboratory. Finally, students will be encouraged to summarize the first part of the project in a document to be presented in an engineering conference.
Faculty:Dr. Rebekah Oulton
Department: CEENVE
Email: roulton@calpoly.edu
Co-Advisor: Dr. Priya Verma, NRES
Accepted Projects Modes: In-Person, Hybrid, Fully Remote
Higher education institutions are increasingly incorporating the concept of sustainability into the curricula, but the evaluation of students’ level of sustainability knowledge is underdeveloped. The proposed research focuses on Cal Poly’s efforts to integrate sustainability principles within the engineering curricula and the impact on students’ knowledge. By analyzing existing data collected on the sustainability literacy of Cal Poly’s engineering students, this project aims to enhance teaching strategies and align them with the University’s sustainability goals. This initiative seeks to better prepare engineering students to become future contributors to sustainable development by effectively embedding sustainability competencies within their educational programs.
Faculty: Dr. Stefan Talke
Department: CEENVE
Email: stalke@calpoly.edu
Co-Advisor: Dr. Serena Lee, CE/ENVE
Accepted Projects Modes:
In-Person
How fast is coastal land sinking? Assessing vertical land motion (VLM) is a surprising—but pervasive– challenge in assessing sea-level rise and the evolution of coastal flood risk, due in large part to cost. Nonetheless, the problem is real; for example, VLM in the Sacramento Delta drives rates of sea-level rise that are greater than 1 cm/year, much larger than oceanic rates (see our recent paper, https://www.nature.com/articles/s41598-023-49204-z ). In this potential SURP project, we plan to investigate and implement inexpensive ways of assessing vertical land motion, using off-the shelf differential GPS (GNSS) systems that have greatly reduced in price over the past decade. The inexpensive GNSS technology will be incorporated into our existing array of inexpensive water level sensors and deployed at coastal docks along the Central Coast. The preliminary goal for summer 2024 will be to assess vertical elevation to within 1 cm, as verified by surveying equipment. The use of software post-processing to reduce uncertainty and bias will be explored. The work is potentially important because it fills an unmet need in the current generation of commercial instrumentation. No water level instruments incorporate GNSS technology, in large part because GNSS sensors have only recently become affordable ($100-1000). The scientific community recognizes the importance of GNSS measurements for projections of relative sea-level rise, since future impact depends on whether land is sinking or rising. Hence, our project will explore the engineering feasibility and scalability of incorporating inexpensive GNSS measurements in our deployment package. Initial results show that we can easily get an accuracy of 10 cm, so the focus will be on improving our approach to get to the needed 1 cm accuracy. Experience with electrical engineering, mechatronics, computer hardware, or programming is preferred. Because the proposed work includes both hardware development and software post-processing, two SURP students would be preferred.
Faculty:Dr. April Grow
Department: CSSE
Email: amgrow@calpoly.edu
Accepted Projects Modes:
In-Person or Hybrid
Embroidery software is heavily proprietary, targeted at professionals with high-end machines. This project aims to capture the utility of embroidery software — controlling an embroidery machine to produce embroidered output — and combine it with the playfulness of a game, to put those advanced controls in the hands of novices. Think Kid Pix for machine embroidery. The project will explore a deep understanding of the physical limitations of machine embroidery, ensuring correctness of patterns in an unobtrusive way, while also applying procedural content generation and artificial intelligence techniques to generate surprising and delightful patterns and designs. This project advances the areas of creative computing and interdisciplinary applied computing, and aims to entice more people, especially women, to become interested in STEM technologies.
Faculty:Dr. Ayaan Kazerouni
Department: CSSE
Email: ayaank@calpoly.edu
Co-Advisor: Dr. Zoë Wood, CSSE
Accepted Projects Modes: In-Person, Hybrid
Computing coursework that focuses on people rather than things is more likely to attract and retain students from demographic groups that have been historically underrepresented in computing. Specifically, many students are more likely to “opt in” to computing when they believe that they will be able to use their burgeoning computing knowledge to benefit their communities, society, and humanity. It is therefore our goal to transform coursework in early computing courses to include a sustained focus on computing for good. To that end, we are working within an NSF-funded alliance of six CSU campuses, with faculty creating programming assignments in which students use programming to address communal problems. In order to scale this effort, we have created an assignment portal where faculty can publish their assignments for others to remix and reuse. However, various CSU campuses teach their introductory courses using different programming languages. In this SURP project, we would like to “”translate”” instructional artefacts (assignments, examples, etc.) between programming languages. For example, at Cal Poly SLO, we have created a number of assignments in languages like TypeScript and Java, which our colleagues at CSU Fullerton would like to adapt and use in C++. This translation will involve writing examples, starter code, reference solutions, and test cases in several programming languages. In many cases, these instructional materials are quarter- or semester-spanning sequences of assignments, requiring a non-trivial effort that is appropriately scoped for a summer project. This effort will aid in our broader goal of expanding participation in computing—as more courses adopt materials that connect computing with society, we expect to see a strengthened sense of belonging in computing, and consequently improved 4- and 6-year retention rates of students who are historically underrepresented in computing. Finally, deploying multiple versions of assignments will give us a unique opportunity to explore and compare novices’ challenges with learning programming in various contexts. The “”same”” assignment offered in different institutions using different programming languages will lead us to an improved understanding of pitfalls and affordances of the programming languages commonly used to teach introductory programming.
Faculty: Dr. Bruce DeBruhl
Department: CSSE
Email: bdebruhl@calpoly.edu
Co-Advisor: BJ Klingenberg
Accepted Projects Modes: In-Person, Hybrid
In this project, students will develop a scalable environment for securely analyzing malware including wild-type and custom-developed samples. This will include students developing a custom malware sample to use as a test case, using cutting edge software analysis tools, and developing custom analysis tools. Overall, students will be expected to work in modern programming languages including Rust, GO, C, and C# as well as ARM64 and AMD64 assembly. As needed, students can learn these languages while working on this project.
Faculty:Dr. Daniel Frishberg
Department: CSSE
Email:dfrishbe@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid, Remote
Prior work—a 2019 study by Schreffler, Vasquez, Chini, et al.—investigated the efficacy of Universal Design for Learning (UDL) in supporting students with disabilities (SWD) in science, technology, engineering, and mathematics education (STEM). SWD are somewhat more likely than the general population to enroll in STEM majors, but for some disabilities, students are less likely to graduate with those majors. This highlights the need for more research on the efficacy of accessibility strategies and frameworks. The authors found a paucity of prior work involving the use of UDL for SWD in STEM. In addition to highlighting the need for future quantitative research overall, the authors noted that it would be interesting to investigate which UDL methods are most effective for SWD. For example, for certain executive functioning disorders or forms of neurodivergence, is it more appropriate to assign a greater number of lower-stakes assessments—providing multiple means to access information—or a smaller number of higher-stakes assessments—reducing the cognitive load of balancing a large number of deliverables? Are think-pair-shares an effective, low-stakes environment in which to discuss ideas, or do they introduce burdensome social pressures?
We aim to investigate this question of which UDL methods are most effective, beginning with the number of assessments, through a combination of data aggregation and qualitative interviews. Specifically, our research questions are:
Research Question 1: How much variability is there in the numbers of summative and formative assessments used in CS courses?
Research Question 2: What impact does a greater number of low-stakes assessments, or a greater ratio of formative to summative assessments, have for neurodivergent students and students with disabilities?
Research Question 3: To what extent do these impacts differ across forms of neurodivergence or disabilities?
We plan to gather data on what assessment patterns are presently used at Cal Poly, from some or all of the following sources: course syllabi from instructor web pages, campus-wide Canvas repository data (if available through CTLT), data on use of accommodations in Cal Poly STEM courses (if available through DRC). This data gathering process will provide us context for asking our research questions.
The next phase, possibly in the term following the SURP project, will be to conduct interviews of faculty, students, or both.
Faculty: Dr. Daniel Frishberg
Department: CSSE
Email:dfrishbe@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid, Remote
The PI is engaged in ongoing work on experiments designed to gather evidence for or against conjectures in one or both of two active areas in theoretical computer science: Markov chain Monte Carlo (MCMC) mixing times, and greedy approximation algorithms for combinatorial optimization problems. Research in the area of MCMC mixing seeks to prove theoretical upper and lower bounds on the mixing times—or convergence times—of random walk-based algorithms for sampling from various distributions. The PI has published three papers, in 2023, on this research, which were entirely theoretical. Two of these papers—comprising the PI’s dissertation—left open a significant gap between upper and lower theoretical bounds, with some intuition in both cases that the lower bound is more likely correct. The PI is currently working with two MS students, who have implemented code to run a Markov chain for sampling binary trees. The PI has also worked previously with undergraduate students at the University of California, Irvine (UCI), implementing a Markov chain for sampling independent sets in trees. Further work extending this research may give evidence for or against theoretical conjectures regarding the mixing time of this chain. A separate theoretical line of research with opportunities for experiments lies within the broad area of approximation algorithms. We focus on the nearest-neighbor chain technique, which uses a stack-based data structure to speed up greedy algorithms. The PI co-authored a 2019 paper on new applications of the nearest-neighbor chain, for speeding up geometric algorithms. A direction the PI and the other authors had hoped to pursue, but have not found new theoretical guarantees, was whether this technique could be used to speed up combinatorial optimization problems, such as finding an approximate maximum independent set in a graph. Implementing this algorithm and running it on random graphs—particularly in the Erdos-Renyi random graph model—may be an interesting direction. The PI plans to work with a Cal Poly student in the spring on a senior project implementing these experiments
Faculty: Dr. Dongfeng (Phoenix) Fang
Department: CSSE
Email: dofang@calpoly.edu
Accepted Projects Modes: Hybrid
The objective of this SURP proposal is to explore how to optimize privacy-preserving while performing data analysis. Privacy-preserving technologies allow models to be trained without having direct access to the raw data and prevent these models from inadvertently storing sensitive information about the data. Exist privacy-preserving technologies include federated learning (FL), differential privacy (DP), homomorphic encryption (HE), and secure multi-party computation (SMPC). These privacy-preserving technologies have limitations and costs. Some of them would be overly burdensome in contexts, while others would be insufficiently secure to protect data and models from the actions of malicious actors. An appropriate combination of techniques for a specific project can only be decided once a proper trust model is built and the various trade-offs of techniques are clearly understood. Therefore, in this project, we will propose different trust models. Healthcare datasets will be used for studying the trade-offs of FL, DP, HE, and SMPC. Furthermore, we will apply the different combinations on the healthcare dataset to observe the cost and performance. This study will offer an ideal reference for data scientists to choose the right tools to optimize performance and preserve patient privacy at the same time.
Faculty:Dr. Foaad Khosmood
Department: CSSE
Email: foaad@calpoly.edu
Co-Advisor: Dr. Cameron Jones, HIST
Accepted Projects Modes: In-Person
Based on existing research we began last year, we would like to continue to explore the lives of African Californios (individuals of African descent who lived and worked in Spanish and Mexican California). The primary mode of exploration is processing texts and records of late 18th and 19th century Spanish records using data science and computational methods. Specifically we would like to study population dynamics, reasons for immigration and emigration to the region and use data visualization techniques to map the footprint of this population on the state of California and beyond. Knowledge of Spanish, data science, and natural language processing is almost necessary and strongly preferred.
Faculty: Dr. Franz Kurfess
Department: CSSE
Email: fkurfess@calpoly.edu
Accepted Projects Modes:
In-Person, Hybrid, Remote
This summer activity is part of an ongoing research project, AI in Search and Rescue, and a continuation of prior SURP projects. Building on the base implementation consisting of a front-end for the paper versions of Search and Rescue forms, a dashboard for the command center, and multiple components incorporating a variety of AI methods, our focus this summer will be on the integration of the AI components into a coherent interface for the command center dashboard. The IntelliSAR Clue Graph is a network graph visualization, inspired by the work on network graphs at the DigiLab, University of Georgia (DigiLab, 2023; GISGeography, 2019; Overcash, 2024). This person-centric graph can be accompanied by additional visualization, such as Point-to-Point Analysis, an Origin-Destination Cost Matrix, Coverage Areas Analysis, or Resource Optimization. We are exploring the use of Generative AI methods (e.g., Large Language Models such as ChatGPT, Google Gemini, Anthropic Claude, and Microsoft CoPilot) to create the visualizations from our data repository.
Faculty: Dr. Franz Kurfess
Department: CSSE
Email: fkurfess@calpoly.edu
Accepted Projects Modes:
In-Person, Hybrid, Remote
As an outcome of the Cal Poly “Big Ideas” Strategic Initiative in 2019/20 a group of faculty worked on an Ethics and Social Justice (ESJ) sub-committee to develop guidelines for project proposals and reports consistent with the motto “Data for Good, Data for All”. The direct outcome yielded a questionnaire, a worksheet, and a collection of materials for workshops, as documented in (Kurfess et al., 2022). The instruments were used in multiple workshops at Cal Poly and Hochschule München, and in several classes by students assessing their own project ideas or completed team projects by other students. Further work by students working mostly with Dr. Kurfess explored the use of AI to support this assessment by analyzing project proposals and related documents concerning the criteria identified in the instruments (Nolan, 2022; Ramchander & Landenberger, 2023). As a SURP 2023 project, Kylene Landenberger and Arya Ramchander incorporated Large Language Models into an approach, with the Cal Poly SLO ESJ Chatbot prototype as the outcome. The results were used for informal experiments in Dr. Kurfess’ CSC 480 Introduction to AI and CSC 580 Graduate AI classes in the F23 and W24 quarters. Related class projects examined the use of LLMs for academic counseling, course registration, and guidance with senior projects.
Specific Aims and Objectives
The main goal of this SURP project is to expand the use of Artificial Intelligence, in particular Large Language Models, to support the assessment of project proposals and similar documents concerning ethics and social justice aspects as identified in the original assessment instruments (Kurfess et al., 2022). In support of this goal, we have the following objectives:
UI Redesign: Revise the user interface for the ESJ Chatbot. In particular, it should allow users to upload documents describing the project to be assessed.
LLM Interface: Abstract the LLM interface so that it can be used with multiple LLMs; currently, it uses only OpenAI’s ChatGPT.
Prompt Engineering: Fine-tune the prompts used as input to the LLM.
ESJ Ontology: Expand and revise the conceptual framework of the ESJ domain.
Repository: Expand the repository of relevant background documents.
Dashboard: Show an overview of recommendations for authors or reviewers of documents, arranged by the ESJ criteria from the assessment instruments.
Visualization: Use radar charts or similar visualization methods to show the scores for one or more projects; this allows for easier comparison.
Applications: Explore the use of the underlying methods and components for senior project selection, course registration and academic counseling
Faculty:Dr. Javier Gonzalez-Sanchez
Department: CSSE
Email: javiergs@calpoly.edu
Accepted Projects Modes:
In-Person, Hybrid, Remote
This project explores the integration of augmented reality (AR) and Emotion AI technologies to enhance user experiences in physical environments. By seamlessly merging virtual elements with real-world contexts, we aim to deepen individuals’ interactions and perceptions of their surroundings. Leveraging AR technology enables users to access contextual information, engage with interactive content, and navigate spaces with heightened immersion and understanding. Additionally, Emotion AI enhances these experiences by detecting and responding to users’ emotional states, fostering personalized and emotionally resonant interactions. We aim to integrate digital content within physical environments using mixed-reality headsets equipped with eye-tracking capabilities and consumer-grade wireless EEG devices. Through innovative methods, we seek to enhance the overall user experience and create dynamic, impactful interactions in cultural, urban, and architectural settings. Furthermore, this initiative aims to enrich the educational journey of students by providing hands-on experience in space design, artificial intelligence for enhancing human interactions, and, ideally, technology-driven connections with cultural heritage. By exploring spatial and affective computing topics, students will gain invaluable insights into the potential of emerging technologies to transform place-based affect-driven interactions for diverse audiences.
Faculty: Dr. Jenny Wang
Department: CSSE
Email: jwang96@calpoly.edu
Accepted Projects Modes:
Co-Advisor: Dr. Xuan Wang, IME
Accepted Projects Modes:
Hybrid
This study explores ethics and trust in human-AI collaboration, focusing on ChatGPT in higher education. It aims to understand trust dynamics and ethical implications, providing insights for responsible AI use in academia. Through empirical analysis, it identifies key trust dimensions and evaluates alignment with ethical standards, offering recommendations for ethical AI utilization in academia.
Faculty: Dr. John Bellardo
Department: CSSE
Email: bellardo@calpoly.edu
Accepted Projects Modes:
Hybrid
The main objective of the SAL-E mission is to grow the Cal Poly CubeSat Lab’s (PolySat) capabilities in rapid spacecraft design, as well as to provide the current generation of students with the opportunity to design, test, integrate, and communicate with a satellite. This summer we will be focusing on a few aspects of the spacecraft design and preparing our test facilities for Fall. Specifically, we will be ordering electronics boards that were designed in Spring, testing the boards as they arrive, developing software to interface with the payloads, begin integrated testing, updating the control system in our thermal vacuum chamber, and testing our vibrations procedures. Students participating in this multidisciplinary SURP will gain hands-on experience with a few aspects of the spacecraft design cycle while helping move SAL-E toward launch.
Faculty:Dr. Jonathan Ventura
Department: CSSE
Email: jventu09@calpoly.edu
Co-Advisor: Dr. Long Wang, CEENVE
Accepted Projects Modes: In-Person, Hybrid, Remote
This research aims to establish a deep learning-based approach for reconstructing microstructures of nanomaterial networks based on microscopic images to facilitate the investigation of the electrical properties of the nanomaterials. In particular, microscopic images of nanomaterial networks under mechanical loading will first be experimentally obtained and used as target images. A pre-trained convolutional neural network will be adopted with customized loss functions to reconstruct microstructural images that are statistically equivalent to the experimental ones.
Faculty: Dr. Jonathan Ventura
Department: CSSE
Email: jventu09@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid, Remote
This research aims to explore deep learning-based methods for improving and analyzing microscope images in multiple contexts. The SURP project will encompass two separate aims: image denoising and computer-aided diagnosis. In image denoising, the challenge is to automatically restore a short-exposure image to the quality of a long-exposure image but with far less harmful light exposure applied to the sample. For computer-aided diagnosis, the challenge is to develop and compare machine learning methods for accurate and explainable computer-aided diagnosis using microscope images of breast tissue. We are collaborating with a microscopy group at the University of Colorado to develop large-scale datasets for these tasks. The denoising dataset contains images of various biological structures imaged with both long and short exposure times, and the breast cancer classification dataset contains images of breast tissue labeled according to the type and stage of cancer present.
Faculty: Dr. Jonathan Ventura
Department: CSSE
Email: jventu09@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid, Remote
The goal of this research project is to enable a casual user to capture a scene with consumer hardware (a smartphone or a consumer 360 degree camera) and then process it so it can be viewed with free movement in a VR headset. For this SURP project the team will build upon a recent project from my lab called 3D Panorama Inpainting (https://jonathanventura.github.io/3d-pano-inpainting) that can automatically convert a 2D panorama into a 3D scene. We will investigate possible paths to improving the results such as incorporating new methods for depth estimation and inpainting, and processing panoramic video to capture dynamic content.
Faculty: Dr. Joydeep Mukherjee
Department: CSSE
Email: jmukherj@calpoly.edu
Accepted Projects Modes: Hybrid, Fully Remote
Cloud services have lately transitioned from a single monolithic application to several loosely coupled microservices. This design shift from a single complex monolithic application to tens or hundreds of single-purpose microservices has enabled modern cloud services to be deployed, scaled, and managed in a more efficient way. Microservices can be deployed inside their own lightweight containers hosted on a cloud platform, and they can be individually developed, updated, tested, and elastically scaled. Each microservice can also be developed in its own suitable language, with only a common API for microservices to communicate with each other. Due to these benefits, large cloud providers and services such as Amazon, Apple, Netflix, and Twitter have adopted the microservice architecture model in recent years. In this project, we will study and deploy a known distributed cloud service benchmark suite that is representative of a real industrial application with tens of microservices hosted on multiple cloud virtual machines. This suite covers a broad spectrum of popular cloud services such as social networking service, media service and an e-commerce service. Each service consists of multiple microservices that use different languages, programming models and open-source frameworks. The microservices will communicate using RESTful APIs and will enable simulating user traffic flow through a load generator tool. The final project outcome will generate a modular and extensible distributed microservice application suite that is deployed on a popular cloud platform and can be used by faculty members and students for performance debugging and simulating a practical representative industrial cloud service.
Faculty:Dr. Ka Yaw Teo
Department: CSSE
Email: kteo@calpoly.edu
Accepted Projects Modes:
In-Person, Remote
Select one of two projects:
Optimizing Sensor Placements for Fixed Source Localization: A Distinct Subset Distance Sum Problem
Imagine that we are given a set P of n points in the open half-plane defined by a line L. Our objective is to locate a set S of k points on L, for some positive integer k, so that for any subset Q of P, there exists at least one point s Î S such that the sum of the (reciprocal) Euclidean distances from Q to s is distinct, thus allowing for a unique identification of Q.
This problem formulation is inspired by a theoretical interpretation of non-invasive brain-computer interfaces (BCIs), which allow us to detect and utilize electrical brain signals to interact with external devices such as computers and robots through sensors placed on (or under) the scalp. To accurately translate brain signals into output commands, we need to identify the (subset of) neural ensembles from which the electrical brain signals are relayed. This leads us to modeling the source (signal origin) localization problem as aforementioned, where P corresponds to the set of neural transmitters emitting electrical signals over a resistive medium (brain matter), L represents the scalp, and S denotes the set of sensors to be optimally placed on the scalp. The direct path between a sensor s Î S and a transmitter p Î P can be seen as an electrical circuit, in which the electrical current is inversely proportional to the distance between p and s. To ensure that a sensor can deduce any possible subset of active transmitters through the sum of their electrical currents, we require that all subset distance sums must be distinct. It is worth noting that only a fraction of the transmitters might be active at any given time.
The applicant proposes a study seeking to address the problem described above and its variants. In particular, the study will investigate the following questions:
Q1) What is the minimal cardinality k* of S required to fulfil the condition of having distinct subset distance sums? Is it always possible to achieve said condition with just k* = 1 point? What happens to k* if we require the absolute difference between any two distinct distance sums to be larger than a given threshold?
Q2) Is there an efficient deterministic algorithm to solve the general problem? If not, what efficient heuristics may be possible?
Q3) What if P lies on L?
Q4) What if S is no longer constrained to L and can be anywhere in the plane?
Q5) What if P and S must be integer points (in an integer lattice)?
Our problem seems (somewhat) related to the construction of a subset sum distinct (SSD) set, which is defined as a set of numbers in which any two subsets have different sums. An SSD set can be easily constructed in linear time either through creating a sequence of increasing numbers where each number in the sequence is greater than the sum of the preceding numbers, or by computing the so-called Conway-Guy sequence [Guy, 1982]. Interestingly, in comparison, our problem appears to be more challenging due to its geometric settings and constraints.
The proposed study is expected to generate theoretical results on a new, non-trivial combinatorial problem in geometry, which could be of interest to computational geometers and practitioners alike. The new findings will help to motivate new algorithmic developments in environment sensing applications by obtaining a better understanding of the problem at hand.
Predicting Tumor Response in Childhood Osteosarcoma: An Application of Geometric and Topological Data Analysis
Radiomics focuses on extracting and analyzing quantitative features from medical images, with the goal of identifying prospective biomarkers that can facilitate the diagnosis, prognosis, and clinical decision-making for various diseases. While advanced machine- or deep-learning-based radiomics has been successful in uncovering implicit valuable features that may not be apparent to human observers, it necessitates an ample quantity of data with similar clinical characteristics to yield reliable results regarding tissue heterogeneity, which forms the basis for solving most image classification and segmentation problems in biomedicine. Additionally, the unexplainable nature of these implicitly extracted features remains a critical issue, hindering the clinical adoption of deep-learning radiomics. Understanding the rationale behind predictive model outcomes is imperative to prevent irreparable loss or injury due to inexplicable mistakes in medical decisions.
These challenges are even more prevalent in pediatric cancer research, particularly when studying rare disorders such as high-grade osteosarcoma — the most common form of bone cancer in children. Our goal is to create a classification model for predicting tumor response to neoadjuvant chemotherapy in pediatric osteosarcoma patients using multi-modal magnetic resonance imaging (MRI). Unlike other tumor types, childhood osteosarcomas exhibit inherent heterogeneous, with significant variability in histology and anatomy, making it problematic to establish a correlation between image intensity profile (or other MRI-derived parameters) and tumor necrosis. In our previous work [Teo et al., 2022], we addressed these impediments by leveraging different texture-based features (i.e., quantitative spatial relationships among pairs of voxels) statistically indicative of tumor necrosis versus viable tumor. These texture features were proven adequate for establishing a proof-of-concept correlation model based on c-means clustering with reasonable accuracy.
Topological and geometric features have shown potential in capturing shapes and local patterns that may be associated with medically observable characteristics. However, no studies have yet been conducted to interpret and correlate these geometric statistics with visual evidence and patterns perceived by medical experts through similarity measures or distance metrics. Such a correlation could help explicate the subsequent statistical predictive model, further justifying its validity alongside empirical evidence. The applicant proposes a study seeking to explore and advance the potential utility of state-of-the-art topological and geometric summary statistics in the medical domain — specifically, in various MRI modalities for diagnosing and monitoring the progress of osteosarcoma. The study is expected to contribute to the advancement of radiomics and improve our understanding of the role of geometric and topological features in biomedical image processing and analysis.
Faculty:Dr. Ka Yaw Teo
Department: CSSE
Email: kteo@calpoly.edu
Accepted Projects Modes:
In-Person, Remote
In minimally invasive robotic surgeries, a needle-like robotic probe is typically inserted through a small incision to reach a given target region. Once in place, the probe can perform various operations such as tissue resection and biopsy. The latest variants of these robotic probes have an articulating joint close to the tip, enabling the surgeon to rotate the tip, if necessary, after inserting the probe in a straight path. This feature allows the surgeon to reach deep and obscure targets effectively but also increases the complexity of finding acceptable insertion-and-rotation combinations for the robotic probe. As with any other robotic arm, this type of robotic manipulator is highly scalable, adaptable, and can be used for other applications beyond the medical domain.
Unlike polygonal linkages that can rotate freely at their joints while moving between a start and target configuration, a simple articulated probe as described above is limited to a fixed sequence of moves, which involves a straight-line insertion followed by a possible rotation of the end link. Motion planning involving such a constrained robotic probe has not been studied rigorously through a geometric lens until recently. Teo et al. [2020] considered the trajectory planning problem for such an articulated probe and provided an efficient deterministic algorithm for finding the set of all feasible “”extremal”” solutions in O(n2 log n) time, where n is the size of the input obstacles. Although inexact heuristics, such as probabilistic roadmap (PRM), rapidly-exploring random tree (RRT), and approximate cell decomposition planners, are generally effective in solving a wide range of motion planning problems, they have been known to incur a high variance in solution quality and running time, especially when planning manipulative tasks for multi-link robots with multiple constraints.
The applicant proposes a study seeking i) to develop approximate planners using sampling- and subdivision-based strategies for solving said trajectory planning problem and assess their empirical performances in comparison to that of the deterministic algorithm, and ii) to extend Teo et al.’s deterministic algorithm to solving the trajectory planning problem involving an articulated probe with k links, where k is any fixed positive integer. The project involves elements from multiple disciplines such as computer science theory, robotics, and computer-assisted medical interventions. The implementations and comparison results will shed light on the practicality, robustness, and limitations of the different approaches as well as providing indicators for potential future work that could be done to optimize and improve the existing planners. The extension to k-link articulated robot will serve as an important step towards achieving an automated planner for finding feasible trajectories for snake-like robotic instruments.
Faculty: Dr. Ka Yaw Teo
Department: CSSE
Email: kteo@calpoly.edu
Accepted Projects Modes:
In-Person, Remote
Consider the following problem: Given a set P of n points in d-dimensional real space, where d is any fixed positive integer, locate a point in P that minimizes the sum of the Euclidean distances between P and the located point. The optimal point for the problem is commonly referred to as the medoid. The medoid problem, by its very nature, is fundamental and interdisciplinary given its relevance and application to a variety of modern-day data analytic tasks in numerous areas ranging from biology and medicine to natural language processing and social science. One would encounter the problem of computing the medoid in various contexts — clustering in data science, facility location optimization in operations research, pattern recognition in computer vision, and centrality quantification in network analysis, just to name a few.
Naively, one can find the medoid of P by simply computing all (n choose 2) pairwise distances. However, it has been argued that an exact algorithm does not exist for solving the medoid problem in o(n2) time [Newling and Fleuret, 2017]. Hence, various randomized approaches have been developed to compute the medoid in sub-quadratic time either approximately or exactly under certain statistical assumptions [Eppstein and Wang, 2004; Feldman and Langberg, 2011; Baharav and Tse, 2019]. Most recently, Daescu and Teo [2022] designed the first deterministic fully polynomial-time approximation schemes (FPTASs), which rely on the principle of well-separated pair decomposition, with a time complexity near-linear in n.
The applicant proposes a study seeking i) to implement and potentially simplify the deterministic algorithms and compare their empirical performances to those of randomized methods, and ii) to extend the deterministic approach of Daescu and Teo’s algorithms to solving the k-medoids problem, which involves finding k medoids that minimize the sum of the distances between points in P and their nearest medoid. The proposed study is expected to produce an extensive experimental analysis of non-trivial algorithmic results, which could help to motivate new algorithmic developments while achieving a better understanding of the problem at hand. In addition, the extension to k-medoids will be a valuable contribution of interest to the theoretical community as well as practitioners.
Faculty: Dr. Ka Yaw Teo
Department: CSSE
Email: kteo@calpoly.edu
Accepted Projects Modes:
In-Person, Remote
As computer technology becomes increasingly integrated into our daily lives and decision-making processes, it is essential to ensure that computer science (CS) education goes beyond technical proficiency and instills in students the ability to make ethical and socially responsible decisions in their professional domain. This can be achieved by contextualizing technical training using ethics and social justice (ESoJ) elements. Specifically, technical assignments should be complemented with ESoJ narratives that align with technical learning objectives, demonstrate human judgments within technical contexts, and emphasize the social consequences of technical decisions.
However, integrating ESoJ into theoretical CS courses such as algorithms and discrete mathematics/structures poses a challenge, as there is currently a lack of instructional resources that do so. While helpful instructional materials are available for introductory non-theory CS courses, generalizing them to create instructions suitable for CS theory-related topics has proven difficult, and it remains to be seen how to include ESoJ in these topics. Therefore, this proposed work aims to address this challenge and respond to the call by Smith et al. [2023] to generate instructional materials that include ESoJ for various computing courses.
The applicant proposes a study seeking i) to incorporate ESoJ concepts into an algorithms course, and ii) to identify and address instructional design challenges in integrating ESoJ into the CS theory curriculum. The study is expected to identify ESoJ contexts that highlight human judgments involved in algorithm design techniques and to create algorithm-design-and-analysis problems that require both technical prowess and ESoJ reasoning. As there is currently no established roadmap for integrating ESoJ elements into algorithms curriculum, the study intends to investigate various strategies for designing ESoJ-incorporated instructional materials. The study will document the design process in detail and explore ways to address any obstacles encountered during the process. By doing so, the study hopes to provide valuable insight and guidance for educators looking to include ESoJ elements in their algorithms courses.
Overall, the proposed effort aims to support and bolster ethics and social justice education in computer science. Additionally, it seeks to make CS theory courses more relatable, socially relevant, and appealing to students. The ultimate goal of this work is to foster a generation of socially responsible computer scientists well-equipped to confidently make ethical decisions.
Faculty:Dr. Kirk Duran
Department: CSSE
Email: kduran02@calpoly.edu
Accepted Projects Modes:
In-Person, Remote
Resume generation with Large Language Models (LLMs) like ChatGPT is a boon for job applications by automating the creation of customized resumes. Despite their advantages, LLMs often produce drafts that are not ready for submission, lacking in detailed alignment with job descriptions and necessitating significant user modification for verification, formatting, and editing. These models may inaccurately inflate or fail to effectively present an applicant’s relevant experience, compromising the resume’s authenticity and impact. This project proposes an approach to enhance LLM-generated resumes by incorporating feedback from Applicant Tracking Systems (ATS). Through analyzing ATS evaluations of resumes, we gain insights into the characteristics of successful documents. This analysis informs the development of precise prompts for ChatGPT, steering it towards producing resumes that not only match ATS criteria but also genuinely reflect the applicant’s qualifications and experiences. This methodology aims to optimize LLM capabilities for generating tailored, ATS-optimized resumes, significantly improving job seekers’ application efficacy.
Faculty:Dr. Maria Pantoja
Department: CSSE
Email:mpanto01@calpoly.edu
Accepted Projects Modes:
Remote
With the advent of foundational deep learning models and advances in high-performance computing, Information Retrieval (IR) has become an increasingly popular application in Artificial Intelligence. Based on information given by a user (for example), AI technology can cross-reference it with the knowledge & information stored in the computer to make informed decisions, powering everything from search engines and question-answering systems (e.g. automated customer service agents) to content recommendation systems and chatbots (Large Language Models such as Google’s Gemini, Microsoft’s Copilot, Anthropic’s Claude and a variety of others will reference external sources to improve the credibility of their output). However, in order to fully take advantage of these IR capabilities, the model requires access to large volumes of data in the form of a vector database (not to mention the computational costs associated with these large models), leading to outrageous fees to uphold it, meaning that this advantage is accessible to only those who can afford these fees (i.e. massive corporations). From here, the question becomes: “How can we represent this information with less space?” Unlike current methods, which look at each piece of information individually to represent, we aim to look at all the information as a whole in our approach (as is done within the human brain). More specifically, our goal in this project is to develop a Graph Hierarchical Representation Learning method that generates structured latent vector spaces capturing the inherent structure of information at varying levels of abstraction, given an unstructured set of information. Through the hierarchical structure, the dimensionality of each piece of knowledge can be reduced significantly such that the total size of the knowledge and data repository will only be a fraction of traditional methods while still representing the same volume of data. Given the nature of this theory, the percentage of the total size reduced will increase with the size of the dataset. Despite this size reduction, exploring the practical efficiency of this theory requires experiments on high-performance computers to perform comparisons. We will initially use resources available at Cal Poly, such as the MPAC lab, as well as cloud computing services. As we gain deeper insights, we will explore the use of additional resources, such as national high-performance computing centers. The scope of the work will vary depending on the backgrounds of students who are selected for the team, but will nonetheless involve experimentation (on multiple types of hardware), exploratory literature surveys, and coordination. As this project emphasizes experimentation and novelty, it may involve students from any college at Cal Poly. We are especially interested in students who have experience working with large datasets, distributed systems, DevOps, or interdisciplinary research.
Faculty: Dr. Sumona Mukhopadhyay
Department: CSSE
Email: mukhopad@calpoly.edu
Accepted Projects Modes: Hybrid
This research project represents a pioneering endeavor to unlock the communicative potential of plants through the integration of cutting-edge technologies and advanced machine learning methodologies. By harnessing data streams derived from sound/audio and imagery captured by Sony’s Spresense high-performance audio sensor and HDR camera, the project aims to help agricultural monitoring practices. By developing a multi-modal deep learning framework tailored for real-time analysis of crop environments, this work seeks to advance knowledge and understanding in both the fields of plant biology and artificial intelligence. By interpreting plant behavior using ML models and enabling proactive interventions based on acoustic and visual cues, this interdisciplinary approach has the potential to transform our understanding of plant responses to environmental stimuli and empower agricultural stakeholders with actionable insights for optimizing crop health and resilience.
Faculty:Dr. Dale Dolan
Department: EE
Email: dsdolan@calpoly.edu
Accepted Projects Modes: Hybrid, but mostly in person, All
Select one of two projects:
Remote Detection of Solar Photovoltaic Tracking Errors through Power Monitoring Algorithms
The Cal Poly Solar farm has been built as a single axis tracking facility with two different types of panels. Both conventional single cell solar panels and twin cell solar cells have been used in its construction. Twin Cell panels typically perform better than their conventional counterparts when shaded by other panels in the row in front of them in a fixed tilt system. However neither module performs well when even a small portion is shaded. This project will access the system API to access data for the field and process to determine areas of concern and different power profiles will be identified that locate and identify inefficiencies in the operation of the system. Project Mode: Hybrid, but mostly in person
LED based Solar Simulator Development and Testing for Validation of Photovoltaic Cells
The Cal Poly Solar Farm on campus provides a unique opportunity to conduct research and data collection on an industry scale solar PV generation facility that is just not available to other Universities. One area in particular is the testing and validation of solar cells and modules in both a laboratory setting and in the field. Our campus does not have the facilities to conduct the laboratory level testing and validation to complement the opportunities that are available with the Cal Poly Solar Farm. This project would aim to begin to develop that capability. We would develop a small-scale solar simulator based on LED technology that would be capable of reproducing the spectrum of the sun as well as having the capability of producing a wide range of spectral compositions. This is an area of industry interest as more typical solar simulators use a variety of different light sources due to the difficulty of reproducing the solar spectrum with what used to be a more limited variety of LED wavelengths. LEDs provide a much more efficient method that alternate light sources and as such are a desirable substitute as well as providing the capability for variable spectral composition.
This project would involve the design and testing of a solar simulator using a variety of discrete LEDs each which would be controlled separately, binned by colour to produce a varying intensity and thus different spectral compositions. This allows a consistent test bed for the testing of solar cells in laboratory settings and comparisons of different PV technologies by students on campus. Project Mode: All
Faculty: Dr. Jason Poon
Department: EE
Email: jasonp@calpoly.edu
Accepted Projects Modes:
In-Person, Hybrid, Remote
To help curb climate change, the United States is rapidly electrifying many sectors of the energy economy, such as transportation (electric vehicles) and home heating (electric heat pumps). Thus, recent years have witnessed an explosion of businesses that want to connect and operate energy resources on the electric grid. However, getting an energy resource connected to the grid is an arduous process that takes years or more, requires specialized technical knowledge, and requires navigating complex and varying regulatory procedures. This project will develop a software platform that leverages AI and high-performance computing to automate and accelerate grid interconnection and energy asset operation. The platform will help energy project developers reduce their interconnection costs by automating aspects of the grid interconnection and asset management processes, including technical feasibility studies, portfolio optimization, and energy arbitrage.
Faculty:Dr. Kun Hua
Department: EE
Email: kuhua@calpoly.edu
Accepted Projects Modes:
In-Person
Road travel safety is always the most important issue in transportation systems. In general, several factors cause road accidents, such as human error, vehicle mechanical failure, roadway limitations (e.g. pavement, lane geometry, etc.), and inclement weather conditions. The major focus of today’s transportation developments is related to making highway transportation safer, smarter, and greener to enhance livability. Many accidents are caused when drivers lack a better understanding of the surrounding traffic conditions because the driver not only needs to control his/her vehicle but also needs to be aware of the movements of the vehicles around him/her. A driver cannot be fully focused on the road continuously due to many factors, such as distractions or fatigue. A mobile wireless sensor network (MWSN) of interconnected vehicles and infrastructure has the potential to both improve traffic flow and increase safety for the traveling public. According to the U.S. Department of Transportation’s Intelligent Transportation Systems (ITS) Joint Office, interconnected vehicle applications provide connectivity between and among vehicles, infrastructure, and wireless devices to prevent crashes, provide safety, mobility, and environmental benefits, and provide continuous real-time connectivity to all users. These systems enable drivers to have 360-degree awareness through vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) connectivity. Collectively V2V and V2I connectivity can be called V2X communication. In addition to providing information to drivers, this connectivity generates an enormous amount of new data about traffic patterns and road usage. ITS can apply this data for both short-term routing decisions and long-term planning for improved safety and reduced congestion. These short- and long-term optimizations at both the vehicle and infrastructure levels will enable vehicles to reach optimal fuel efficiency. Travel safety can be improved if vehicles can be designed to form groups for mutual interaction with each other to form a MWSN through V2X communications for intelligent transportation systems (ITS). With an increasing capacity of sensors available on vehicles, a need has arisen to make collaborative measurements of the information collected by individual vehicles to form an enhanced data set to improve the inter-vehicle safety features. Vehicles of the future will communicate both with each other and with the infrastructure to better serve drivers with autonomous maneuvering, traffic data for improved navigation and route optimization, analytics for increased traffic flow and safety, and automatic accident reporting for reduced emergency response times.
Faculty:Dr. Kun Hua
Department: EE
Email: kuhua@calpoly.edu
Accepted Projects Modes: Hybrid
Faculty: Dr. Payam Nayeri
Department: EE
Email: pnayeri@calpoly.edu
Accepted Projects Modes:
Hybrid
Antenna measurement systems are used to measure the radiation characteristics of antennas across their operating frequency band. Given the high cost of microwave components, commercial microwave antenna measurement systems are extremely expensive. As a result, most undergraduate students never gain experience in how to design an antenna measurement system. This research aims to change that by creating a systematic process to design and develop a low-cost platform for antenna measurement systems, that can be used by any student to design and build their own system.
These systems generally consist of a vector network analyzer (VNA), a rotary machine, a reference antenna, multiple cables and connectors, and an antenna-under-test (AUT) which the system will characterize. The most expensive components are generally the VNA and the reference antenna. Low cost options for VNA are now available such as Seesii NanoVNA-F V3 which operates up to 6 GHz. Commercial reference antennas, however, are still very expensive.
The reference antenna is a key component of these systems and needs to create a very precise linearly polarized electromagnetic plane wave that will be used to characterize the AUT. Most commercial antenna measurement systems use multiple standard gain horn antennas in order to cover the frequency range of interest. These horn antennas however are limited by the range of frequencies supported by the waveguide feed, thus requiring multiple horns for wideband measurements. A broadband antenna will eliminate this issue and also reduce system costs. However, there are multiple other aspects to consider.
This undergraduate research aims to create a single wideband reference antenna that can operate from 1 to 6 GHz for a low-cost antenna measurement system. This frequency range covers most of the Wi-Fi bands, and 5G/6G sub-6GHz bands, making it an extremely practical measurement system. While multiple types of wideband antennas exist, linearity, gain, polarization, dispersion, and size/weight are other important factors to consider for a reference antenna. This research aims to investigate different types of antennas that are capable of such wideband operation as a reference antenna. One potential solution is a discone antenna, however the student researcher will investigate other potential options. This research is a hands-on experience. So, in addition to learning the fundamentals of antenna radiation, analysis, and design, the student will also gain knowledge in CAD modeling, fabrication, and experimental antenna testing. The prototype that the student will develop will be fully characterized in the existing anechoic antenna test chamber, and later used as the reference antenna for the proposed low-cost antenna measurement system.
It should be noted that the development of the low-cost antenna measurement system is a student senior project, starting in Fall 2024, however, analysis and design of the system reference antenna is a key research need, and is pursued through this undergraduate research.
Faculty: Dr. Payam Nayeri
Department: EE
Email: pnayeri@calpoly.edu
Accepted Projects Modes:
Hybrid
Faculty: Dr. Siavash Farzan
Department: EE
Email: sfarzan@calpoly.edu
Accepted Projects Modes:
In-Person, Hybrid
Select one of two projects:
Advanced Localization and Grasping Sensor Technologies for Autonomous Robotic Apple Harvesting
This research seeks to revolutionize apple harvesting through the development of innovative sensor technologies, advanced computing systems, and automation solutions. Our objective is to create a robotic system capable of efficient, accurate navigation within complex orchard environments, reliable apple detection under varying lighting and occlusion, and precise grasping with minimal fruit or tree damage.
This project will focus on several key technical aspects. Firstly, it will implement sensor fusion techniques (e.g., Kalman Filtering) to combine GPS and encoder data, thereby optimizing robot navigation accuracy and mitigating the limitations of individual sensors. Secondly, it will design a perception system utilizing vision cameras and depth sensors to reliably identify apples under diverse lighting conditions and amidst obstacles, ensuring accurate and selective harvesting. Thirdly, the project will integrate haptic and force sensors into the robotic gripper, enabling real-time measurement and adjustment of grasping force to minimize fruit bruising while maintaining a secure hold. Finally, a central computing unit will be used to facilitate seamless communication between localization, perception, and grasping subsystems, ensuring coordinated movement and adaptive, damage-free harvesting.
Extensive testing in both controlled and real-world orchard environments will validate the system’s robustness and adaptability. This project aims to significantly advance sensing and automation within agriculture, leading to greater efficiency and yield while safeguarding product quality.
Collective Localization and Dynamic Formation Control for UAV Swarms
This research project aims to develop advanced distributed computing algorithms and adaptive control strategies for multi-agent Unmanned Aerial Vehicle (UAV) swarms, enabling robust operation in complex environments. The core focus is on cooperative localization, where UAVs leverage inter-agent communication and sensor fusion techniques (e.g., Extended Kalman Filters) to improve positioning accuracy, particularly in GPS-denied environments. This enhanced localization forms the foundation for developing dynamic formation control algorithms. These algorithms will allow swarms to optimize task performance and obstacle avoidance by adaptively reconfiguring their formations in response to real-time environmental conditions and mission requirements. The project combines theoretical algorithm development with validation through simulation and hardware experiments, ultimately aiming to revolutionize the autonomy and efficiency of UAV swarm operations across various applications.
Faculty: Dr. Siavash Farzan
Department: EE
Email: sfarzan@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid
This research project aims to develop a groundbreaking educational curriculum and associated hardware prototyping resources for “”Drive to Succeed””, a DEI initiative designed to introduce autonomous driving concepts to underserved high school students. Two undergraduate engineering students will collaborate with faculty to design a curriculum that integrates theoretical and practical aspects of autonomous systems. Students will gain hands-on research experience in hardware prototyping, sensor integration, and developing effective teaching materials that resonate with diverse learners. The project emphasizes a modular approach, allowing students to explore core concepts like motion planning, control, perception, and embedded systems in a structured manner. This initiative directly addresses underrepresentation in engineering fields and seeks to empower a new generation of engineers from all backgrounds.
Faculty: Dr. Duha Ali
Department: IME
Email: duali@calpoly.edu
Co-Advisor: Dr. Javier Gonzalez-Sanchez, CSSE and Dr. Rafael Guerra-Silva, OCOB
Accepted Projects Modes: In-Person
Emotions are a crucial factor in communication, affecting the quality and effectiveness of interactions. As AI technology continues to advance, there is a growing interest in discovering how these entities elicit and respond to human emotions. This project aims to explore the emotional response in human interactions with AI entities, utilizing the advanced capabilities of brain-computer interfaces.
Through this project, we seek to compare emotional responses elicited during conversations with AI counterparts to those in interactions with fellow humans. Utilizing a brain-computer interface (Emotiv MN8 headset), which offers high-fidelity emotional data collection, we will capture and analyze various emotional markers such as arousal, valence, and engagement.
The methodology involves designing controlled conversational scenarios in which participants engage with AI-enabled systems and human counterparts. The Emotiv MN8 devices record participants’ emotional responses in real time, providing rich insights into the dynamics of these interactions.
By examining the emotional nuances in human-AI conversations, this project aims to shed light on how individuals perceive and engage with AI entities on an emotional level. The findings hold implications for designing and developing AI systems that aim to foster more emotionally resonant user interactions, ultimately contributing to advancing human-centered AI technologies.
Faculty: Dr. Jill Speece
Department: IME
Email: jespeece@calpoly.edu
Accepted Projects Modes: Hybrid
Western University Medical School receives a staggering 10,000 applications annually for only 300 coveted positions, necessitating a refined selection process to identify top-tier candidates eligible for interviews with the highest potential to become successful medical school graduates. A successful student achieves cumulative board scores of 99% and secures their #1 choice residency placement. Despite this influx, a significant portion of applicants lack essential information, hindering their progress through the enrollment pipeline. This project aims to augment the pool of interview-eligible candidates through data-driven methodologies and predictive analytics.
The student researcher will collaborate closely with the Associate Vice President of WesternU Oregon, Di Lacey, DHSc, and their faculty mentor to analyze historical data and develop a robust screening mechanism to identify promising candidates for interviews. They will have access to two sets of data from the past decade – enrollment data and student outcome data – using them to discern application indicators associated with ultimately successful students to inform current selection criteria. The developed screening mechanism will be utilized on existing applications, and the outcomes will be compared with the current selection method of screening by committee. The project will also address issues like “”melt,”” where a student submits an application, accepts, then gets a better offer and leaves.
Ultimately, this project aspires to revolutionize medical school enrollment by integrating advanced predictive analytics into the selection process, ensuring Western University Medical School consistently enrolls the most qualified and promising individuals, thereby ensuring excellence in medical education and practice.
Faculty: Dr. Mohamed Awwad
Department: IME
Email: mawwad@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid
The resurgence of manufacturing reshoring in the United States has significant implications for supply chain resilience and sustainability. This research project aims to develop a comprehensive framework using supply chain digital twins to assess the impact of manufacturing reshoring on supply chain dynamics, performance, and sustainability. By leveraging the capabilities of anyLogistix software and Business Intelligence tools, the project will create a virtual representation of the supply chain to simulate, analyze, and optimize reshoring scenarios. The research will explore the economic, environmental, and operational benefits and challenges of reshoring, providing valuable insights for policymakers, industry leaders, and supply chain professionals. The outcome will contribute to the development of resilient and sustainable supply chains in the era of Industry 5.0, aligning with national strategies for revitalizing American manufacturing and securing critical supply chains.
Faculty: Dr. Puneet Agarwal
Department: IME
Email: pagarw05@calpoly.edu
Accepted Projects Modes: Hybrid, Remote
In this project, we will explore a range of factors that influence the salaries of US college graduates. Demographic variables such as gender, race, and ethnicity; educational variables such as fields of study; and professional variables, including job and employer characteristics, will be examined to understand their impact on salaries. Our research will extend to a spatial and temporal analysis, offering a deeper understanding of how these factors vary across different states and over time, revealing critical disparities linked to gender, race, and ethnicity. The study will provide valuable insights that can help individuals make well-informed decisions regarding their career trajectory, assist advocates for workplace equality in their efforts to promote fairness, and guide employers towards addressing wage gaps to ensure equal pay for equal work. The project will leverage the extensive dataset provided by the National Survey of College Graduates (NSCG), a biennial survey that has been gathering detailed information on college graduates since the 1970s. The NSCG is sponsored by the National Science Foundation and conducted by the Census Bureau. The project will embrace the principles of the Teacher-Scholar model, offering a hands-on learning experience for student researchers. Through active engagement in data collection, processing, and analysis, along with the utilization of advanced statistical and data analytical techniques, students will gain both theoretical knowledge and practical skills. To translate our findings into actionable insights for decision-makers, policy analysts, and the general public, we will create an interactive dashboard using Tableau/PowerBI. This tool will offer a dynamic and user-friendly interface for the exploration and interpretation of our research findings.
Faculty: Dr. Xuan Wang
Department: IME
Email: xwang12@calpoly.edu
Co-Advisor: Dr. Jenny Wang, CSSE
Accepted Projects Modes:
Remote
Food waste poses a significant environmental and economic challenge to our society. Traditional ways to manage food inventory have several major drawbacks, such as inaccurate inventory counting, inefficiency to count inventory in a timely matter, and labor-intensive work to analyze the inventory. Therefore, there is an urgent need to modernize inventory management. The Shelf Intelligence project aims to tackle these issues by leveraging the power of Artificial Intelligence (AI) to automate inventory counting and management process. A computer vision system utilizing existing hardware (smartphones, tablets, and cameras) will be the platform for AI software development. The software will capture the image information of each object, categorize the objects, create digital twins, and store the digital twins in a database. A dashboard of inventory will be visualized to provide users an intuitive view and assist the decision-making process to dispose the inventory according to the recommendations by the software.
Faculty: Dr. Seamus Jones
Department: IME
Email: sjones97@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid
Aqueous organic redox flow batteries (AORFBs) are energy storage devices with potential to produce a cost-effective solution to current energy storage needs associated with decarbonization of our domestic energy supply. Although this technology offers a great opportunity, the design and selection of cost-effective, water-soluble organic molecules with high stability and suitable redox potential has not been optimized. Recent studies by Jeff Law and coworkers have used machine-learning to identify suitable compounds in terms of stability, synthesizability, and redox potential, however the candidate molecules have not been analyzed in terms of their aqueous solubility, which ultimately controls the energy density of the device. We will investigate aqueous solubility of several organic compounds first by using practical solubility experiments to develop an initial training dataset. Secondly, we will use this training dataset to identify effective models to examine aqueous solubility for these classes or organic molecules. Finally we will apply these effective models to expand the work of Jeff Law and coworkers to account for ideal electrolyte candidates with appropriate consideration to aqueous solubility. These candidate molecules will be synthesized and tested in a Cal Poly senior project, continuing the work beyond the SURP period.
Faculty: Dr. Thale Smith
Department: MATE
Email: tsmith50@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid
Refractory high entropy alloys (RHEA) are high melting point metallic materials comprised of five or more principal elements. The microstructural implications of alloying with yttrium as a principal element for RHEA materials are not well understood. This project aims to use computational thermodynamic modeling to test the microstructure sensitivity of RHEA materials systems to alloying concentrations reported in literature in order to identify criteria for effective microstructural prediction of yttrium containing RHEA materials. A list of promising yttrium containing RHEA will be generated for future experimental validation of identified criteria. This sensitivity analysis will refine and enhance computational models to more accurately predict microstructures crucial for designing RHEA materials.
Faculty: Dr. Alan Zhang
Department: ME
Email: zhangas@calpoly.edu
Accepted Projects Modes: In -Person
Tensegrity structures are composed of stiff rods and elastic cables suspended in a flexible tension network. In particular, the biotensegrity model proposes that all biological systems exhibit tensegrity-like characteristics across multiple scales, ranging from the cellular level to the musculoskeletal system of tendons, ligaments, and fascia, to the human body as a whole. Compared to traditional biomechanical models, it can be a more accurate representation of how motion emerges from natural forms, but further work is needed to fully understand the heterarchical nature of human anatomy. This project will focus on the design and modeling of a tensegrity shoulder joint, in the context of assistive devices to augment human dexterity in the upper limb. The goal is to add shoulder support functionality to the upper limb exoskeleton prototype. Preliminary literature review suggests the six-bar icosahedron as a fundamental building block for the hierarchical nature of the tensegrity system. The SURP student will further explore suitable tensegrity configurations through rapid prototyping and integrate the shoulder with existing tensegrity upper limb designs.
Faculty: Dr. Amanda Emberley
Department: ME
Email: acemberl@calpoly.edu
Co-Advisor: Dr. Ben Lutz, ME
Accepted Projects Modes: Hybrid, Remote
Sociotechnical thinking is an essential part of engineering and requires engineers to consider and design for the impacts of their decisions on the environment and society. However, foundational level courses are often taught with decontextualized problems and examples. The purpose of this project is to incorporate modules focused on sociotechnical thinking into engineering statics and dynamics courses. These modules will help students incorporate both technical and non-technical factors into an engineering solution. We have developed initial drafts of these modules. For example, one of the modules asks students to analyze the impact of levee design on the Pajaro River in Watsonville, CA as a result of the 2023 flooding. The purpose of this SURP project is 1) to analyze the initial implementation of these modules in terms of their impact on student learning and engineering identity and 2) iterate and refine them for use by broad audiences of engineering students. The work will entail qualitative analysis of student responses to the initial implementation of the modules, and we will leverage that analysis to refine and update the modules in ways that make them more effective and interesting. This SURP will serve as the foundation for a larger, collaborative NSF proposal in which the PIs will scale these efforts to additional universities in and community colleges throughout CA. If funded, there is potential for the SURP student to continue work on the project after summer 2024.
Faculty: Dr. Behnam Ghalamchi
Department: ME
Email: bghalamc@calpoly.edu
Accepted Projects Modes: Hybrid
This project aims to explore the integration of augmented reality (AR) technology into engineering education, particularly within Cal Poly’s laboratories, with the goal of enhancing learning outcomes and student engagement. Augmented reality offers immersive and interactive learning experiences by overlaying computer-simulated objects onto the physical environment. A key challenge we aim to address is achieving real-time simulation, which we plan to accomplish using the Speed-goat real-time machine to ensure smooth simulation operation. Our methodological approach will involve phased implementation, including iterative testing and optimization cycles, along with systematic deployment of the Speed-goat real-time machine and calibration procedures.
By harnessing AR technology, students will gain the ability to visualize complex engineering concepts, simulate real-world scenarios, and interact with 3D models in a more intuitive and engaging manner. Through qualitative and quantitative assessments, the project will evaluate the effectiveness of AR-enhanced learning activities, with the ultimate goal of advancing innovative pedagogical approaches in engineering education.
Faculty: Dr. Eltahry Elghandour
Department: ME
Email: eelghand@calpoly.edu
Accepted Projects Modes: Hybrid
Continuous fiber reinforced 3D printing is an advanced additive manufacturing technique for producing polymer matrix composites (PMCs). This method takes advantage of the desirable mechanical properties of composite materials while reducing restrictions on design imposed by traditional manufacturing techniques. This project seeks to characterize, design, and test Kevlar reinforced structures produced via the Markforged platform in an effort to better understand the behavior of additively manufactured PMCs under out-of-plane impact loading. The students involved in this project will work to optimize the 3D printing process for mechanical performance, investigate the effects of the design parameters (fiber volume fraction, fiber reinforcement strategy, and infill pattern) on impact strength & failure mode, and contribute to practical testing of the structures via drop weight impact testing. Lastly, the project aims to develop finite element models to evaluate the composite’s structural performance and validate them with experimental data.
Faculty: Dr. Eric Ocegueda
Department: ME
Email: ocegueda@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid
Granular materials are ubiquitous in nature; their presence spans geo-mechanical fields (in the form of gravel and sand), agricultural fields (as grains or crops), pharmaceutical fields (in the packaging of pills), and environmental fields (the discrete ice structures formed by the melting of ice caps). In each of these fields, ’failures,’ such as the flow of soil (leading to landslides) or jamming of grains as they funnel into containers, are difficult to predict due to the multi-scale behavior of granular materials. The discrete nature of these materials leads to complex bulk behavior dependent on the local physics of particles in crucial areas. The current literature has taken phenomenological or multi-scale approaches to capture these complexities. In phenomenological approaches, we assume the existence of explicit stress-strain equations, allowing for efficient simulations at the cost of uncertainty from the ’simple’ equations. For multi-scale approaches, simulations of individual grains are coupled with bulk-scale simulations, leading to more certainty at the cost of computation time. The proposed project aims to address the computational cost of multi-scale models by using machine learning to develop surrogate models. We will train a neural network on computational stress-strain data of granular materials to construct an accurate replacement of discrete-scale simulations. Thus, two SURP students will work on connected sub-projects: (1) generating computational data of granular materials through packaged software and (2) training a neural network using pre-obtained data (and eventually data from (1)).
Faculty: Dr. Eric Ocegueda
Department: ME
Email: ocegueda@calpoly.edu
Accepted Projects Modes: In-Person, Hybrid
Equity gaps in higher education are a long-recognized issue, with demographics such as race, ethnicity, and socioeconomic class closely tied to percentages of college attendance, completion, and success. Although recent CSU-wide initiatives have aimed at closing these equity gaps, Cal Poly’s unique position among the CSU campuses – as the only campus transitioning into a Hispanic-serving institution – necessitates increased efforts on our campus. In addition, increased equity gaps in STEM fields and Cal Poly’s strong engineering program require a focus on retention and academic preparation of underrepresented minorities (URM) in engineering fields. Thus, the proposed project aims to explore common academic challenges URM students face in engineering and propose structured supplemental courses to address observed issues. In the first portion of the project, we will work with the College of Engineering dean’s office to access student success/demographic data and conduct student surveys to find first-year courses with pronounced equity gaps (likely introductory mathematics or physics classes). Secondly, we will create outlines of supplement workshops for one of the recognized courses to promote academic and college readiness through culturally informed instruction and mentorship from URM engineering faculty. By creating an internal engineering workshop (and future implementation), we will contrast previous programs by taking an engineering perspective to train engineering students in math/physics, allowing motivation/connection of material to future engineering courses. We are seeking one student interested in diversity, equity, and inclusion efforts on campus to aid in the creation of such course workshops. Students may consider highlighting in their application how their various identities might positively impact what they bring to the research project.
Faculty: Dr. Hans Mayer
Department: ME
Email:hmayer@calpoly.edu
Accepted Projects Modes: In-Person
The breakup of the edge of a liquid sheet, i.e., the transformation of a bulk liquid from a planar sheet into small droplets – similar to what can occur in the formation of a spray, is a fundamental fluid dynamics question that I explored experimentally as a PhD student. At that time, I was attempting to experimentally verify a lingering question as to the instability mechanism that results in the breakup of the liquid sheet edge. Those experiments involved creating soap films bounded by metallic wire edges, and then detaching those soap films by rapid boiling along the wire frame. The detached soap film edges could then retract, the edge become unstable, and eventually break up into droplets. Experimental results from both my work, and another work published in the literature at around the same time, were inconclusive and a reasonable hypothesis was that the experiments lacked control over the initial spatial disturbance along the edge of the film (along the unadulterated wires used). In order to truly gain insight into the growth rate of the instability along the edge, a method of creating known spatial disturbances is necessary and the challenge of doing this is the length scales involved are those of the thicknesses of soap films, i.e., at the microscale. Thus, as part of my funded years long project, we are attempting to develop the experimental capability to create patterned soap film frames using traditional microfabrication techniques. By ditching metallic wires and instead patterning silicon wafers, we will be able to achieve the initial spatial disturbance control we seek, and therefore gain more conclusive insight into the instability that grows at the edge of a retracting liquid sheet. The SURP project will play an integral role in helping to continue with the development of the experimental equipment and procedures necessary for performing these fundamental fluid dynamics experiments. This is one of many extra-curricular and curricular undergraduate student research/project opportunities (e.g., it was included in addition to funding students in OUDI’s BEACoN program and using an ME Senior Project team) that were specifically outlined in the NSF-ERI proposal which is funding the project and making this type of exploration at Cal Poly SLO possible.
Faculty: Dr. Hyeonik Song
Department: ME
Email: hsong13@calpoly.edu
Accepted Projects Modes: In-Person
Trade Space Exploration (TSE) is a decision-making process used to evaluate various solutions and their trade-offs to select the most preferred solution. It is widely used in engineering and design as it facilitates the systematic and quantitative exploration of various options. It reduces human bias and uncertainties in the selection of optimal design or system. Computational models for TSE enable the use of tools to automate some aspects of the process which includes analysis and visualization. However, many TSE tools rely on human decision-making for thorough and timely exploration. Understanding the appropriate timing and methods for human intervention in computational TSE is essential. In this summer research project, we aim to develop a human-in-the-loop computational TSE tool to tackle a complex and practical problem: forming ME 428 senior design project teams from a pool of over 150 students. The complexity of this problem arises from the need to consider various selection criteria, including students’ preferences for specific projects, their desired team members, their working style, and the constraint that teams must consist of four students. Currently, the selection process is manual and relies on a trial-and-error approach, which can take hours to achieve an appropriate combination. The TSE tool developed in this project will be used to support the team formation process. We plan to test the efficacy of the tool in the coming Fall quarter of 2024.
Faculty: Dr. Hyeonik Song
Department: ME
Email: hsong13@calpoly.edu
Accepted Projects Modes: In-Person
The practice of self-repairing and maintaining of one’s own device, ranging from small electronics to large appliances, significantly contributes to e-waste reduction. As the global demand grows for incorporating repairability into product design, in line with the right-to-repair movement, it’s crucial to examine individuals’ experiences with repairing their own devices and explore ways to integrate self-repair into engineering education. The objective of this summer research project is to understand the participation and impact of self-repair practices among students from diverse backgrounds, with a particular focus on those traditionally underserved in STEM. To advance this objective, we will undertake two tasks in this project. Survey: We will use survey data from students at Cal Poly as well as from visitors of Cal Poly Repair Café to identify potential barriers that hinder underserved students in STEM from self-repairing their devices. We will apply Natural Language Processing (NLP) and qualitative data analysis methods to identify themes within the survey data, which then can be used to develop recommendations on how to improve self-repairing processes. Additionally, we will explore potential correlations between self-repairing practices and factors such as socio-economic background, prior technical knowledge, and perceived barriers to repair. Literature Review: We will conduct a systematic literature review to compile recommendations for broadening the participation of underserved students in STEM education, particularly in extracurricular settings such as university makerspaces. We will identify opportunities and challenges in implementing these practices within the context of self-repair, such as in university repair cafes. The insights gained from these tasks will inform recommendations for improving the equity in self-repair experience and create learning opportunities for diverse students, such as workshops and curriculum integration, that enhance their abilities and engagement with the right-to-repair movements.
Faculty: Dr. Leily Majidi
Department: ME
Email: lmajidi@calpoly.edu
Accepted Projects Modes: In-Person
Shape memory polymers (SMPs) represent a category of smart materials that can be programmed into a temporary shape and undergo a viscoelastic strain recovery to return to their permanent shape upon exposure to appropriate stimuli, such as heat, electricity, light, magnetic fields, and solvents. The actuating behavior of SMPs promotes them to become promising candidates in various applications such as soft robotics, microfluidics, biomedical, and electronic devices. The SMP research has advanced significantly in the last decade, yet there is a lack of comprehensive research on the response of SMPs to multiple stimuli. In this summer’s research project, two students will focus on boosting the shape recovery degree and efficiency of SMPs using low-cost fabrication methods and polymer composites.
Faculty: Dr. Melinda Keller
Department: ME
Email:mkeller@calpoly.edu
Accepted Projects Modes: In-Person
In 2022 aviation accounted for 2% of global energy-related CO2 emissions, having grown faster in recent decades than rail, road or shipping. Help Cal Poly partner with Boeing to investigate cryogenic fuels as an alternative energy source for commercial aircraft to bring down CO2 emissions. This project will entail setting up a continued space for these kinds of projects at Cal Poly, studying the state of the art in cryogenic storage and use, and doing a feasibility study for your ideas on cryogenic propulsion proposals. Focus will be on developing standard operating procedures and a safety mindset for continued projects with this industry partner.
Faculty: Dr. Mohammad H Hasan
Department: ME
Email: mhasan04@calpoly.edu
Accepted Projects Modes: In-Person
In this proposed work, the implementation of vision and hearing through machine learning will be pursued. Vision and hearing will be implemented on newly acquired Quanser Q-Arm robotic systems, which are available in the mechanical engineering robotics lab. The Q-Arms acquired have their own proprietary software which interfaces with MATLAB, named QUARC. As such the goal of this project is two-fold: (1) Produce vision and auditory models using available software libraries like TensorFlow. (2) Implementing the vision and auditory models developed on a physical robot arm (Q-Arm) through a MATLAB interface.
Faculty: Dr. Ramanan Sritharan
Department: ME
Email: rsrithar@calpoly.edu
Accepted Projects Modes: In-Person
At present the applications and various material properties of Carbon Nanotubes (CNTs) have been extensively studied. CNTs have excellent mechanical, electrical, optical, thermal, and magnetic properties. CNTs are broadly classified into two major types, i.e., straight CNTs (SCNTs) and helical CNTs (HCNTs). The potential application of CNTs is their use to improve the mechanical properties of structures in hi-tech applications such as airframe structures and wind turbine blades. Epoxy polymer resin is one of the widely matrix systems to fabricate polymer-based composite structures. Drop-weight tests and open-hole tension tests are some of the widely used mechanical tests in industries to characterize materials. In this research project, epoxy polymer resin will be reinforced with different configurations of CNTs with various weight percentages. Samples will be prepared and tested based on ASTM D7136/D7136M or ASTM D5766/D5766M standards. The primary objective of this research is to investigate the effect of different configurations and weight percentages of CNTs in epoxy polymer resin through the mechanical test.
Faculty: Dr. Ramanan Sritharan
Department: ME
Email: rsrithar@calpoly.edu
Accepted Projects Modes: In-Person