MIT AeroAstro Northrup Grumman Undergraduate Research and Innovation Scholar
Designing and Implementing Geometric Reasoning Solvers for RMPL Planners with a PMT Approach
Brian C. Williams
I propose to design a PMT-based language for geometric reasoning in RMPL to be integrated in path planning algorithms for an Autonomous Underwater Vehicle. Employing the formal techniques of PMT will require developing core and module files for the solver investigating the effectiveness of automatic heuristic generation with a geometric reasoning-based domain and implementing the solver that an RMPL planner will call. Exploring the development of one PMT-based solver will hopefully lead the way breaking down all solvers into core and module files to be used by the planner. Successful implementation of the geometric reasoning solver will allow for the eventual transition of the planner and sub-solvers to one that fully implement formal Planning Modulo Theories.
I am an upcoming junior studying courses 16 and 6-2. I am personally interested in the domains of Autonomous Systems, Robotics, and Space Exploration. Im excited to work on this SuperUROP because of its applications to many different types of rovers, spacecraft, and autonomous vehicles. I hope to gain experience with AI with respect to robotics, and plan to use what I learn to continue working in space exploration!