Research On Key Technologies Of Multi-UAV Cooperative Planning And Control

Unmanned aerial vehicles(UAVs)have been widely applied in military and civil fields due to their rapid development.The ability of a single UAV is limited by its payload and endurance,while cooperation of multiple UAVs can enhance the task efficiency and enrich the mission functionality.Thus,multi-UAV cooperation are attracting significant interests from both theoretical and application research.With the hardware foundation provided by the advanced computers,sensors and communication networks,the cooperative planning and control methods,which currently are the technical bottleneck for multi-UAV cooperative applications,are critical for the performance of multi-UAV cooperation.Multi-UAV cooperative planning and control is subjected to multiple complex constraints including mission requirements,flight environment and UAVs maneuverability,and it must consider the tight couplings among the time,space and tasks of different UAVs.Based on the hierarchical framework,the key technologies of multi-UAV cooperation are studied,which include cooperative task assignment,cooperative path planning,cooperative trajectory planning and cooperative tracking control.Moreover,the multi-UAV indoor cooperative flight testbed is established to verify the effectiveness of proposed cooperative planning and control methods.The multi-UAV cooperative task assignment with heterogeneous targets is studied for cooperative reconnaissance problem.The heterogeneous targets are classified into point targets,line targets and area targets according to the features of target geometry.Given the kinematic constraints of UAVs and the objective function of reconnaissance task,the multi-UAV reconnaissance task assignment for heterogeneous targets is formulated as an extended multiple Dubins travelling salesmen problem.Considering the characteristics of task assignment constraints,a modified opposition-based genetic algorithm(OGA)is developed to improve the population variety and enhance the global exploration capability,where the double-chromosomes encoding and multiple mutation operators are tailored.The simulation results demonstrate that the modified OGA outperforms ordinary OGA and random search for the multi-UAV cooperative reconnaissance task allocation problem.To trade the optimality and efficiency of multi-UAV path planning,cooperative path planning is divided into preplanning and dynamic planning.Path preplanning prefers to improving the optimality of results.Within the coupled cooperative planning scheme,the cooperative path preplanning is formulated as a constrained nonlinear optimization problem.And a constrained particle swarm optimization using filter mechanism is proposed to avoid using penalty function and to improve the feasibility and robustness of the results.Dynamic path planning prefers to ensuring the planning efficiency.Based on the successive cooperative planning scheme and hierarchical planning ideas,a given-range-constrained sparse A* search(SAS)algorithm and a height planning algorithm are developed to achieve fast paths planning.Simulation results show that the proposed methods can provide satisfied paths for multi-UAV cooperation.Considering the objective of UAV tasks and constraints of temporal and spatial cooperation,minimum-time trajectory planning for multi-UAV cooperation is formulated and solved by sequential convex programming(SCP).Multi-UAV cooperative trajectory planning is an optimal control problem with nonlinear dynamics and nonconvex path constraints.Using the direct collocation method,the nonconvex optimal control is transcribed into a nonconvex optimization problem.Then,the successive convexification is introduced to reduce the computational complexity of the optimization problem.Through linearization of dynamics and convexification of obstacle avoidance and collision avoidance,the nonconvex problem is converted to be a series of convex optimization subproblems,which are efficiently solved by primal-dual interior-point algorithm.Furthermore,a parallel SCP method is proposed to improve the efficiency and enhance the scalability to the number of UAVs.In parallel SCP,each subproblem is decomposed into multiple independent problems which can be solved in parallel,by decoupling collision avoidance and ensuring time consistency.Numerical experiments verify the effectiveness,optimality and efficiency of the SCP-based cooperative trajectory planning method.Simulation results demonstrate that SCP provides better efficiency than the pseudospectral method and the efficiency advantage of SCP becomes more obvious as UAV scale grows.Multi-UAV trajectory tracking control method is conducted to satisfy the constraints of temporal and spatial cooperative tasks.The six-degree-of-freedom motion equations for UAVs are firstly established.By decoupling and linearizing of motion equations,the transfer functions of longitudinal motion and lateral motion are derived.And the autopilot is designed to control velocity,heading and flight-path angle.Then,the guidance model of UAVs with autopilot in the loop is built,and the guidance law for trajectory tracking is designed based on sliding mode control.Numerical simulations demonstrate that the designed autopilot and guidance law can effectively track the reference trajectory and achieve UAVs cooperation.Besides,cooperative trajectory tracking results indicate the feasibility of planned trajectory and effectiveness of multi-UAV cooperative framework.To further validate the effectiveness of multi-UAV cooperative planning and control methods,multi-UAV indoor cooperative flight testbed is established.Firstly,the overall scheme of the testbed is designed,which includes its compositions,data relations and the performance of main modules.Then,multi-UAV cooperative flight scenarios for typical missions,i.e.,formation rendezvous,formation reconfiguration and formation exchange are formulated.Finally,cooperative flight tests are conducted in the established testbed.The flight test results demonstrate the effectiveness of the developed multi-UAV cooperative framework and the proposed cooperative planning and control methods.