Task Placement Optimization of Redundant Robots Using Reliability Maps Under Locked Joint Failures

Abstract

The failure-tolerant workspace of a robot is defined as the reachable workspace both before and after an arbitrary joint is locked at an arbitrary angle during a motion. Once the task is located within the failure-tolerant workspace, task completion can be guaranteed. However, a failure-tolerant workspace is typically only a small region within the robot workspace, and existing methods are limited to computing it in the positional workspace. This work introduces reliability maps, which provide information on the reliability of different regions across the entire workspace. A method is developed to compute reliability maps in both positional and orientational workspace accurately and efficiently. The computed reliability map can be applied to optimize end-effector task placement to maximize the task completion ratio. Examples of reliability maps for a planar 3R robot, a spatial 4R robot, and a spatial 7R robot are computed using the proposed method, along with uniform and random sampling for comparison. Simulation and physical experiments are conducted to further validate the proposed task placement optimization method. Results indicate that using the reliability map to optimize task placement can significantly improve the task completion ratio after experiencing arbitrary locked joint failures.