Control Software of Robot Compliant Wrist System

The compliant wrist combining passive compliants and sensor has been developed in GRASP laboratory. The device provides the robot system the necessary flexibility which accommodates transitions as the robot makes contact with the environment, corrects positioning error in automatic assembly, avoids high impact forces and protects the surface from damage. The device also supplies the displacement sensing of the passive compliance so that active feedback control is possible. This report is intended to serve as a reference material to introduce the control software of the robot compliant wrist system developed and implemented in the lab. The detail discussion on system performance and parameters selection can be found in the thesis [3]. The rest of material is organized as follows. Section 2 introduces the compliance control methods of robot manipulators. The historic development of both passive and active compliance method is discussed. The advantages and disadvantages of the methods are investigated. Based on the unsolved problems in this issue, the six-degree freedom compliant wrist is developed, and the design feature is presented. Section 3 discusses the hybrid position/force control scheme using the sensing information from the device. The positioning error due to load or external force when robot moves in free space is compensated for, so that the effective stiffness is increased. In force control when robot is constrained by environment, the trajectory is modified by sensed force, so that the effective stiffness is decreased. Section 4 deals with the implementation of the control scheme. Various programs have been developed to perform the hybrid control operations, such as hybrid control demonstration, surface tracking, edge tracking, insertion and pulling out, and writing operation. The programs have been successfully implemented in the experiments. Definition and selection of the parameters in the programs are discussed. Section 5. is the source code of control scheme which has been implemented in PUMA 560 with index machine in GRASP Laboratory. The control is executed on a MicroVax I1 using the RCI primitives of RCCL