Structure-Specified <i>H</i> _∞ Loop Shaping Control for Balancing of Bicycle Robots: A Particle Swarm Optimization Approach

Abstract

The Particle swarm optimization (PSO) approach is an efficient meta-heuristic search method that can be used to solve the multi-objective and non-convex optimization problems that are normally generated when designing structure-specified H ∞ loop shaping controllers. In this study, PSO is used to design the controllers able to balance a bicycle robot, with a model-based systematic procedure for the controller design being proposed. Simulation and experimental results are presented that show that superior robustness and efficiency properties were obtained using the structure-specified H ∞ loop shaping controllers compared to those obtained using the proportional plus derivative as well as conventional H ∞ loop shaping ones. In comparison with the solutions obtained based on genetic algorithms, the use of PSO has a better efficiency in term of the computational time.