Leighton Collins, IV | Robotics and Software Engineering

QLBM is a novel multi-robot task planning algorithm using iterative clustering of a discretized space, to partition a given area into sections to be assigned to each robot within a group of robots. This partitioning method accounts for "no-fly" zones, as well as the robots states, which ensures a more even load distribution amongst the robots, to provide a lower total time required to complete each mission. A nearest neighbor path planning algorithm is then used to reorder the list of cells assigned to each robot to optimize the path taken for each robot, in a computationally efficient manner. This algorithm was developed as a result of working on DARPA's 2019 OFFSET Program, and became a major portion of my thesis contribution. In Chapter Three of my thesis, I explain the need for such an algorithm, its uses, how it works, and present test cases and results. This chapter is provided below for further reading. The algorithm proves to be scalable to hundreds of robots and large areas with a linear trend in computation time with respect to the number of robots.