Distributed Cooperative Onboard Planning for the Conflict Resolution of Unmanned Aerial Vehicles

Abstract

The distributed cooperative onboard conflict resolution problem of unmanned aerial vehicles is studied in this paper. First, the dynamics of unmanned aerial vehicles and the safe separation constraint are studied. The monotonicity feature of the constraint condition is analyzed. The distributed conflict resolution problem is modeled as a distributed constraint optimization problem. Second, the distributed constraint optimization problem is solved in two steps. The feasible solutions are searched first, and the overall costs are minimized thereafter. The distributed stochastic gradient descent method is proposed to search the local near-optimal solutions. The self-improving mechanism is proposed to enhance the utilization efficiency of data exchanged among unmanned aerial vehicles. The “virtue adjustment” theorem is defined to motivate the cooperative coordination between unmanned aerial vehicles. Simulation results show that, compared with existing short-term algorithms, our algorithm could reduce the summation of additional flight distances and impacts on flight plans of unmanned aerial vehicles significantly and coordinate the flights of unmanned aerial vehicles in complex scenarios.