Hierarchical Probabilistic Estimation of Robot Reachable Workspace

Abstract

Estimating a robot’s reachable workspace is a fundamental problem in robotics. For simple kinematic chains within an empty environment this computation can be relatively straightforward. For mobile kinematic structures and cluttered environments, the problem becomes more challenging. An efficient probabilistic method for workspace estimation is developed by applying a hierarchical strategy and developing extensions to a probabilistic motion planner. Rather than treating each of the degrees of freedom (DOFs) ‘equally’, a hierarchical representation is used to maximize the volume of the robot’s workspace that is identified as reachable for each probe of the environment. Experiments with a simulated mobile manipulator demonstrate that the hierarchical approach is an effective alternative to the use of an estimation process based on the use of a traditional probabilistic planner.