A novel, neuro-muscular, actuator control scheme

This thesis presents the design and analysis of a novel actuator trajectory control scheme that draws on the strategies used by the human central nervous system (CNS) to deal with constraints. The hypothesis is that, while the manipulator moves closely along the desired path, it may take advantage of the constraint (if it exists) to minimize its own power output to the environment. A muscle-like model of the actuator is developed in order to demonstrate the motion, while the controller is structured in a manner similar to the human CNS with both an alpha and a gamma command path. The function of the alpha global task control is accomplished by a proportional plus integral (PI) controller, whereas the local stability function of the gamma system is emulated with a novel perturbation controller. The perturbation controller minimizes work output by continuously testing the response of the system to an application of a minute force pulse. The single degree of freedom systems are modelled using MATLAB...