Have a UROP opening you would like to submit?
Please fill out the form.
Developing Safe and Scalable Multi-agent Motion Planners for Large-Scale Multi-agent Systems
16: Aeronautics and Astronautics
Brian C. Williams
As soon as possible
Jingkai Chen: firstname.lastname@example.org
This project is to develop a multi-agent motion planner to coordinate a large group of robots (e.g., ground vehicles and quadrotors). Recent advances in robotics have laid the foundation for building such large-scale multi-agent systems in manufacturing, warehouse logistics, and smart restaurants. A great deal of research has focused on coordinating agents to fulfill different types of tasks. The fundamental task for building these systems, which is also the problem we focus on in this project, is to navigate agent groups from their initial states to goal states while avoiding collisions. The MIT MERS research group recently developed a state-of-the-art multi-agent motion planner to plan agents with nonholonomic, nonlinear dynamics and bounded disturbances by combining control theory, search, and mathematical optimization. In this project, the UROP participant will work closely alongside a senior Ph.D. student to systematically explore the related literature, develop a more powerful multi-agent motion planner based on the existing work, conduct experiments, and summarize the results as technical papers for publications. We will start with the problem involving ground vehicles and quadrotors, and move on to deal with robotic arms. We can also provide the testbed for real-world experiments.
● (Basic) Experience in programming c++ ● (Basic) Familiar with basic heuristic search algorithms ● (Preferred) Excellent ability in technical writing.