Fuzzy Maneuvering Controller Optimization Using an Optimum Differential-Drive Mobile Robot Model

This paper proposes a method for improving the performance of a fuzzy controller applied to a Differential-Drive Mobile Robot. From previously recorded data, the DDMR model is improved through Particle Swarm Optimization aiming to obtain a better representation of the real system. The PSO algorithm
is also applied to adjust the fuzzy controller parameters so that trajectory tracking error is minimized. The manually adjusted controller settings are compared to the optimized's in terms of Root Mean Square Error of trajectory tracking, control effort and acceleration. The latter is useful for protecting the robot from damage caused by abrupt variations. Numerical simulations show that the optimized model can describe better the real system behavior and the optimized controller
can lead the robot dynamic model to track the given trajectory more accurately, with reduced control effort, and smoother velocity variation than the non-optimized controller.