Trajectory Tracking Control of Differential Drive Mobile Robots using Neural Network Disturbance Compensator

Many control methods for trajectory tracking of mobile robots are based on dynamic control strategies, such as computed torque control using coupled nonlinear nominal models in the form of Euler-Lagrange equations. In this paper, we propose a novel control approach to compensate the unknown dynamics of differential drive mobile robots and achieve trajectory tracking using a double integrator-type nominal model and a neural network equivalent disturbance compensator. In the proposed approach, instead of the coupled nonlinear nominal model used in the conventional dynamic control schemes, we introduce a simple double integrator (DI)-type decoupled linear nominal model and treat all the unmodeled dynamics different from that as lumped equivalent disturbances. By estimating and compensating this equivalent disturbance by using neural network, the differential drive mobile robots are resulted in a double integrator-type decoupled linear system and then a control law for trajectory tracking is derived by applying the linear system theory to this system. The effectiveness of the proposed approach is verified through simulations and experiments that are carried out on a four-wheeled differential drive mobile robots.