Automated transport of multiple particles using optical tweezers requires the use of motion planning to move them simultaneously while avoiding collisions amongst themselves and with randomly moving obstacles. This paper develops a decoupled and prioritized stochastic dynamic programming based motion planning framework by sequentially applying a partially observable Markov decision process algorithm on every particle that needs to be transported. An iterative version of a maximum bipartite graph matching algorithm is used to assign given goal locations to such particles. The algorithm for individual particle transport is validated using silica beads in a holographic tweezer set-up. Once the individual plans are computed, a three-step method consisting of clustering, classification, and branch and bound optimization is employed to determine the final collision-free paths. Simulation results in the form of sample trajectories and performance characterization plots are presented to illustrate the usefulness of the developed approach.
Motion Planning for Continuous-Time Stochastic Processes: A Dynamic Programming Approach
