The objective of this dissertation was to develop a real-time computer control system for a Prototype Leg, which is a test bed for evaluating new mechanical designs for a large hydraulically-powered hexapod machine called the Adaptive Suspension Vehicle.; Real-time control was implemented using an Intel 8086 microprocessor-based microcomputer. State Feedback Control with a simplified Inverse Plant feedforward term has been implemented to improve the joint servo performance of high-speed leg motion and to solve the highly underdampled system problem. In particular, filtered pressure feedback is used to speed up the response and improve the damping characteristics.; The new approaches based on the Jacobian have been implemented as the basis for finding the joint positions associated with a constrained foot position. These are numerical approaches and are alternatives to deriving a complex closed-form solution.; It is necessary to control position and force simultaneously during Prototype Leg motion. Modifications of Hybrid Control, which has been applied to robot manipulator control for the purpose of combining both position and force commands in task related coordinates, have been successfully applied to the Prototype Leg and conform to rather stringent computational requirements.; Two proximity sensors mounted on the Prototype Leg have been successfully used for the purpose of stepping over an object during the leg transfer phase. The control algorithm developed is of general form so as to be appropriate for other tasks and configurations of sensors.
展开▼
