Distributed Constrained NMPC with Infeasibility Handling Applied to Formation Control of Nonholonomic Vehicles

This paper presents a distributed constrained nonlinear model predictive controller with infeasibility handling for formation control of a group of nonholonomic vehicles. In this approach, the controllers are distributed to the agents, and the coupling is done explicitly through local objective functions. A strategy based on constraints relaxing is proposed to solve the problem of high-value state and formation errors, factors that directly affect feasibility and stability. Experimental results with commercial robots are provided to demonstrate the performance of the proposed control strategy.