PASO: An Integrated, Scalable PSO-based Optimization Framework for Hyper-Redundant Manipulator Path Planning and Inverse Kinematics

Hyper-redundant manipulation is the use of a hyper-DOF robotic system to accomplish tasks such as picking, placing, reaching, and exploring in challenging environments. Hyper-redundant manipulation involves both finding collisionfree configuration-space paths (the path planning (PP) problem) and transforming a given workspace target pose into a configuration space goal (the inverse kinematics (IK) problem). Traditional, Jacobian-based IK approaches become computationally infeasible as the number of DOF increases. Existing hyper-DOF manipulator PP approaches either require extensive global preprocessing (e.g., Probabilistic RoadMaps) or spend considerable time on costly nearest neighbor computations (e.g., Rapidlyexploring Random Trees). We propose an integrated, scalable optimization framework called PASO (PAth planning with Swarm Optimization) that uses Particle Swarm Optimization in a divide-and-conquer fashion to efficiently produce approximate solutions to both the hyper-DOF PP and IK problems together. Implicitly-defined C-space waypoints are used to bias C-space sampling. We present promising experimental results showing PASO’s ability to scale to much larger manipulators (120 DOF, so far) than reported in the literature on current integrated position and orientation IK and PP solvers for general serial hyper-DOF manipulators in 3D workspaces.