Distributed near optimal flocking control for multiple Unmanned Aircraft Systems

In this paper, we study the distributed optimal flocking control problem of multiple Unmanned Aircraft Systems (Multi-UAS). Using the emerging Neuro Dynamic Programming (NDP) technique, a novel distributed near optimal flocking design is proposed for ensuring the multi-UAS to follow the three heuristic flocking rules (i.e. cohesion, separation and alignment) in an optimal manner. First, an innovative cost function is developed by incorporating system cohesion, separation and alignment performance together. Subsequently, a novel neural network (NN) is proposed to approximate the minimized cost function value by using Hamilton-Jacobi-Bellman (HJB) equation in an online and forward in time manner. Moreover, the near optimal flocking can be attained by minimizing estimated cost function. Using standard Lyapunov stability analysis, the uniformly ultimately boundedness (UUB) of the closed-loop multi-UAS is verified. Simulation results demonstrate the effectiveness of proposed scheme.