Research On Multi-UAVs Formation Flight And Collision Avoidance

With the rapid development of unmanned aerial vehicle(UAV)technology and the more and more complex requirements of flight mission,multi-UAVs cooperative formation flight(CFF)has gradually become a vital research topic.In CFF,the improvement of autonomous collision avoidance ability is of great significance to adapt to more complex environment and improve survivability.Focused on neighborhood collision avoidance for UAVs,several methods to avoid collision are researched in this paper.First of all,the basic concepts and methods for collision detection are discussed.The collision detection and resolution system are introduced.With the elements of UAVs collision avoidance considered,a reasonable mathematical model is established.Based on these works,the velocity obstacle(VO)method is applied to realize real-time collision detection.The corresponding avoidance rules are defined according to the geometric analysis results for different collision threat.Combined with heading adjustment,velocity and height control strategy,a reasonable collision avoidance system is designed.Next,with the direction adjustment approach in collision avoidance,the artificial potential field method is used to calculate the real-time avoidance signal.A reasonable potential function is designed,according which the single-step optimal control signals are calculated.The feasible signal is chosen with the dynamic constraints.Then the approach is extended.Combined with multi-step state prediction,the correctness and validity of the avoidance signal are improved.The effectiveness of the proposed approach is validated with numeric simulations.Finally,an UAV formation simulation with leader-follower pattern is designed,which contains formation keeping and formation reconfiguration for three UAVs.Real-time trajectory guidance method is used to keep the formation.Dubins path planning is applied to reconfigure the formation.Geometric rules and potential field are combined to avoid collisions.The effectiveness of the proposed conflict avoidance method in formation flight is validated by experiment.