Flocking of multi-agent system using a unified optimal control approach

In this paper, the multi-agent flocking problem is investigated in a unified optimal control framework. The desired flocking characteristics, such as velocity alignment, navigation, cohesion, and collision/obstacle avoidance, are achieved by formulating them into respective cost function terms. The resultant non-quadratic cost function poses a challenging optimal control problem. A novel inverse optimal control strategy is adopted to derive an analytical optimal control law. The optimality and asymptotic stability are proved and the distributed feedback control law only requires local information to achieve the flocking behaviors. Various simulation scenarios are used to demonstrate the effectiveness of the optimal flocking algorithm.