Adaptive Control of Robot Manipulators in Varying Environments

The application of autonomous robots is drawing increasing attention in many fields. An autonomous robot can accomplish dangerous or tedious tasks that are difficult for humans, including aerial and underwater tasks. To achieve the task goals precisely and steadily, using a suitable control scheme is vital to an autonomous robot. Adaptive control, with its advantage in overcoming parametric uncertainties, is widely accepted as an advanced control method for robots.

Most research in this field has been focused on the adaptive control of underwater robots, but robot control in a varying environment remains an open problem. This thesis proposes a multiple-model-based adaptive control scheme to deal with the effect of the varying environment. The equations of motion of a robot manipulator are specifically derived. The complete model is based on the original model of a robot manipulator, with the effects of added mass, buoyancy, damping, drag, and lift considered as the varying environmental factors. Then a multiple-model-based control scheme is adopted to deal with the varying environment. Multiple controllers are being compared while controlling the robot. The controller with the slightest error is adopted to compensate for the varying environment effect, and the control performance can be better than a single-model controller. The control scheme is applicable for the robot moving in the air or any other fluid environments. Simulation results of the developed multiple-model adaptive controller on a planar elbow robot moving in the air and underwater are given to illustrate the improved control performance.