Research On Homing Method Of The Parafoil Airdrop System Under Uncertainty

The parafoil airdrop system with excellent gliding and control performance can realize long-distance,accurate and safe delivery,so it has been broadly applied in military and civilian fields.However,the research of flight under uncertainty still faced great challenges,which restricted its further application.In this paper,the problem related to the homing of parafoil airdrop system under uncertainty is studied,including the dynamic model,multi-phase homing method and optimal control method.(1)A 6-dof dynamic model of the parafoil airdrop system is established,with which the gliding and turning features of parafoil are analyzed.For the homing trajectory design,a 4-dof model is established based on the assumption of steady flight,and the key parameters are obtained by data fitting with simulation results from the6-dof model.(2)The T-approach homing method of the parafoil system is studied.A guidance strategy based on waypoints and a proportional controller are designed.The guidance module and control module are connected by the dynamic model to establish a complete flight simulation system.In addition,the control energy consumption with height redundancy is improved by reducing the number of waypoints and adjusting the turning radius of nominal trajectories.In order to calculate the parameters of nominal trajectory,a rapid calculation method of trajectory and an optimization method of the initial waypoint structure are proposed,in which the homing trajectory is simplified to combinations of straight lines and arcs.The flight simulation tests are carried out under uncertainty and the applicability of the proposed method is verified.(3)The PCE method is introduced to the parafoil flight uncertainty analysis,and the influence of different model orders and sample points on the calculation results is analyzed.The calculation efficiency and accuracy are verified by comparison with the Monte Carlo results.The homing problem with no-fly zone constraints is transformed into an optimal control problem,and an optimization model under uncertainty is established with reliability demands.A reliability analysis method based on the PCE model is proposed to ensure that the optimized solution meets the reliability requirements.The above optimal problem is solved by GPM,and the reliability of the uncertainty optimization results is verified.(4)Aiming at the real-time problem of optimal control homing method,a controller combining the global optimal ability of GPM and the learning ability of DNN is proposed.Considering the influence of different network depths,node numbers and learning rates on the training effect,a DNN model suitable for predicting the optimal control of the parafoil system is given.The control prediction accuracy and the homing ability of the proposed method are verified by simulation tests.The study in this paper could provide references to the research of parafoil airdrop system homing under uncertainty with multi-phase homing method and optimal homing method.