Collision-avoidance algorithms are crucial for safe, reliable motion planning. Although most existing algorithms require considerable computation time to randomly generate and verify potential paths, my project will focus on a newly developed and optimized real-time method that generates known collision-free regions and achieves collision avoidance by finding paths only within these “safe” regions. Until now, this method has relied on axis-aligned bounding boxes as the safe regions in order to take advantage of their simple geometric constraints for faster computation. Our research will evaluate the time efficiency of expanding this method to work with more general convex polytopes, which would enable it to be used in a greater variety of environments as well as in the configuration space.

I have really enjoyed the constant learning and discovery that has come with doing UROPs in the past, so I am excited to add to those experiences with the structure and mentorship of the SuperUROP program. I look forward to exploring a different facet of robotics research with the Robot Locomotion Group and hope to gain both publishable results and a more complete understanding of the academic research cycle.