Meet PolyGPT, our AI-powered Project Idea Generator
Tell PolyGPT your favorite subject and a few hobbies you're passionate about to generate personalized project ideas.
Log in to explore PolyGPTExplore project ideas with the help of PolyGPT
See your personalized project ideas come to life with PolyGPT! Log in to brainstorm ideas.
We know the jump from your interests to a research project idea can be tricky.
That's why we've built tools to help you find the perfect project idea to lay the foundation of a successful project. Simply copy your favorite idea to your clip board, and paste it in the project idea section of the application.
Browse project ideas by Polygence mentors
Modern Technology's Impacts on the Environment and the Effectiveness of Renewable Energy Systems
In this project, I mentored a student in exploring how different methods of electricity production affect the environment and how we can reduce carbon dioxide (CO₂) emissions. We investigated renewable energy technologies, carbon capture methods, and emerging solutions for reducing greenhouse gases. A key part of the project was comparing the environmental benefits, costs, and energy requirements of different approaches to understand which solutions are the most practical and sustainable. This project is a great opportunity for students interested in climate change, renewable energy, and engineering to explore real-world environmental challenges, analyze scientific research, and identify promising technologies that could help build a cleaner future.
AI/ML, Engineering, Computer Science, Business
Developing a Novel Robot Localization Algorithm Using Low-Cost Sensors
In this project, I mentored a student who found a simpler way for robots to track their location using everyday low-cost sensors. Traditional methods often need careful measurements of wheel positions and still struggle with sideways drift. My student showed that for certain wheel setups those measurements aren’t necessary, then built a clean arc-based algorithm that reconstructs the robot’s path while naturally correcting for drift. The results were strong enough that we turned the work into a research article, which has been submitted to a journal for publication.
AI/ML, Engineering, Computer Science, Business
The Psychology of Everyday Decisions: Understanding Human Behavior Through Research
Why do people follow social norms even when they disagree with them? How does social media influence self-perception? Why do people procrastinate, make impulsive decisions, or struggle to prioritize their own well-being? Everyday human behavior is full of interesting psychological questions, and this project gives you the opportunity to investigate one of them. You'll explore established theories in cognitive psychology, social psychology, and behavioral science to understand how people make decisions, form attitudes, respond to social pressure, and interact with others. Potential topics include conformity, persuasion, social identity, cognitive biases, motivation, moral judgments, self-control, and the psychology of health and well-being. Depending on your research question, you may conduct a literature-based investigation, analyze existing datasets, or develop a survey or experimental study to explore a psychological phenomenon. I'll help you formulate testable hypotheses, develop an appropriate research design, analyze findings, and communicate your results through an academic research paper. This project is ideal for students interested in psychology, neuroscience, sociology, behavioral science, and understanding why people think and behave the way they do.
Statistics, Psychology, Business
Modeling a Flexible Surgical Needle That Bends Toward Its Target
Many surgeries and biopsies take place deep inside the body, where a straight, rigid needle simply cannot reach without damaging healthy tissue along the way. This project explores how steerable needles and catheters solve that problem: by taking advantage of an asymmetric bevel tip and carefully timed rotation, a flexible instrument naturally curves as it advances, allowing it to arc around obstacles toward a target. Working in surgical robotics and medical device mechanics, the student will investigate how factors such as bevel angle, insertion speed, rotation pattern, and tissue stiffness influence the curvature of the needle's path and the accuracy with which it reaches its target. The student will learn Python programming, the basics of mechanics and geometric modeling, and how to run controlled simulation experiments, varying one parameter at a time and recording how the trajectory responds. The process begins with a survey of image-guided surgery and needle-steering research, moves through building and validating a simulation model against published experimental data, and finishes with an analysis of which design and control choices most improve targeting accuracy. The final outcomes are a working trajectory simulation, a quantitative study of the parameters that govern steering, and a scientific research paper presenting the findings.
AI/ML, Computer Science
Computer Vision for Autonomous Obstacle Avoidance
A drone flying on its own has no pilot's eyes, so it must build its understanding of the world from cameras and learn to react in fractions of a second. This project explores how computer vision allows a drone to perceive its surroundings and navigate safely without human control. Working in aerial robotics and machine perception, the student will study how a drone can detect and track obstacles, estimate its own motion from video, and make simple decisions such as stopping, detouring, or following a target. The student will learn Python programming, the fundamentals of image processing and object detection, and how to train and evaluate a small computer-vision model on aerial imagery, then test it inside a drone simulator where crashes are free. The process begins with a survey of how autonomous drones sense and plan, continues with building and benchmarking a detection model, and culminates in integrating that model into a simulated flight in which the drone must navigate a course while avoiding obstacles. The final outcomes are a working simulated autonomous flight demo, quantitative results on detection accuracy and navigation success, and a scientific research paper presenting the full study.
AI/ML, Computer Science
Designing a Tendon-Driven Prosthetic Hand: How Cable Routing Shapes Grip Strength and Motion
Modern prosthetic hands must solve a difficult trade-off: they need to be strong enough to hold everyday objects, gentle enough not to crush them, light enough to wear all day, and simple enough to be affordable. This project explores how tendon-driven mechanisms, in which cables pulled by a small motor flex the fingers much like the tendons in your own hand, can meet those demands. Working in prosthetics and biomechatronics, the student will investigate how design choices such as tendon routing, pulley placement, and joint stiffness affect a prosthetic finger's range of motion, fingertip force, and ability to conform to objects of different shapes. The student will learn foundational mechanics and kinematics, Python programming, and how to build and run physics simulations of a tendon-driven finger, systematically varying one design parameter at a time and measuring the results. The process begins with a survey of how natural and prosthetic hands achieve grip, moves into modeling and simulation experiments, and, if time and resources allow, can extend to a simple 3D-printed prototype finger tested against the simulation's predictions. The final outcomes are a quantitative comparison of competing designs, a working simulation (and possibly a physical prototype), and a scientific research paper presenting the design study.
AI/ML, Computer Science
Teaching a Simulated Robot Arm to Learn from Demonstration
Robots are remarkably good at repeating exact motions and remarkably bad at handling the messy variability of the real world, and this project explores how learning from human examples can close that gap. Working in robot learning and imitation learning, the student will use Python and the MuJoCo physics simulator to study how a robot arm can acquire a manipulation skill, such as picking up and placing objects, from recorded demonstrations rather than from hand-written rules. Along the way, the student will learn to program a simulated robot, design a simple teleoperation interface, collect and clean demonstration data, train a small neural network policy, and evaluate it rigorously by measuring success rates and cataloguing how and why it fails. The student will begin by surveying recent research on imitation learning, then move through data collection, model training, and iterative testing, keeping a research log throughout. The final outcomes are a working simulation demo in which the robot performs the learned skill, a set of quantitative results comparing performance across conditions, and a scientific research paper presenting the full study.
AI/ML, Computer Science
Structural Analysis of Combustion Chamber Designs in High-Pressure Liquid Bipropellant Rocket Engines
This study uses finite element analysis (FEA) to evaluate the structural integrity of spherical combustion chambers for high-pressure liquid bipropellant rocket engines. Simulations assess stress and displacement across varying design parameters and dimensions to support design optimization. The findings aim to guide manufacturers toward more reliable, efficient chambers and enable comparisons with real-world designs.
Engineering, Physics
Designing Health Reminders People Want to Use
Can a reminder be helpful without becoming another notification people ignore? Choose a focused routine, such as taking activity breaks or preparing questions for a medical appointment, and explore research on reminders and behavior change. We will compare approaches to timing, wording, personalization, and user control, then design a small app mock-up or notification prototype. You will create a simulated evaluation plan that considers usability, privacy, and notification fatigue rather than assuming reminders improve health. Your final product could be an interactive prototype and research-based design report. Coding is optional, and the project can be tailored to your experience. Potential outcomes: Interactive app mock-up or coded prototype; research-based design report; science fair presentation; usability study proposal.
Surgery, Medicine
When Children Outgrow Medical Devices: Designing for Growth
How can an assistive device accommodate a child's changing size and daily needs? Choose one noninvasive device component, such as an adjustable handle or mounting system, and investigate the design challenges through published literature and existing products. We will define requirements, compare adjustment mechanisms, and create a CAD model or physical mock-up. You can evaluate range of adjustment, ease of assembly, and material cost with bench tests. Your final deliverable will be a design report and prototype explaining tradeoffs and next steps. Prototypes are for demonstration, with no patient testing or claims of clinical effectiveness. Potential outcomes: Biodesign prototype or CAD model; bench-testing report; design portfolio; science fair project.
Surgery, Medicine