Neuro-adaptive virtual leader based formation control of multi-unmanned ground vehicles

Abstract

The problem of formation control of multiple unmanned ground vehicles (UGVs) is studied in this paper. The underlying problem is essentially a high dimensional multi-vehicle trajectory tracking control problem, its complexity grows significantly as the number of the involved vehicles increases. To address the singularity inherent in regular formation process and the complexity involved in formation of multiple vehicles, this paper attempts an approach that integrates robust adaptive neural network (NN) control with the virtual leader concept. The fundamental idea behind this approach is to use the virtual leader-follower format to convert the problem into the one that only involves each individual vehicle to track the virtual leader. By doing so, the singularity associated with the traditional formation control is also avoided. As the system involves significant nonlinearities and uncertainties, robust adaptive NN control algorithms are developed, which are shown to be able to achieve high precision formation. The feasibility and effectiveness of the proposed method are also verified by simulation studies.