Swarm Optimized Simple Adaptive Controller for Spacecraft Proximity Operations Trajectory Tracking

Abstract

Abstract Adaptive control design allows for the management of systems with time varying or unknown dynamics. Despite their versatility, few well defned design techniques exist for some classes of adaptive controller. Without analytical techniques it is difcult to prove the efcacy of an adaptive controller design. One solution to this issue is the application of parametric search techniques to adaptive controller design. This paper explores the application of diferential evolution on the simple adaptive control law formulation and compares its solution to one found using particle swarm optimization. Afterwards, variations on these techniques, namely the selection particle swarm optimization and self-adaptive diferential evolution, are implemented and their results compared. The final swarm-optimized controller is compared to a classical Linear Quadratic Regulator (LQR) controller, and a manually designed simple adaptive controller for precision trajectory tracking control of spacecraft proximity operations. Parametric search techniques are able to determine controller parameters that produce a superior control response. Swarm-optimization techniques determine controllers with parameters drastically different from manually designed efforts.