| Modern agriculture needs efficient,accurate and environmentally friendly agricultural production mode,and the rapid development and maturity of UAV technology provides technical guarantee for the application of UAV for plant protection.The realization of intelligent operation of plant protection UAV can improve the efficiency of plant protection operation,reduce labor cost,reduce the use of pesticides and guarantee the quality of agricultural products,which is conducive to the realization of sustainable agricultural development.As a new type of plant protection aircraft,linear plant protection UAV can effectively improve the efficiency of plant protection operation due to its "one-line" fuselage structure,but the unique control model also makes it difficult to fully apply the traditional trajectory optimization and tracking control algorithm on this moving platform.Therefore,this paper studies the trajectory optimization and tracking control methods of linear plant protection UAV,proposes an improved Minimum-snap trajectory optimization algorithm,designs a multi-modal trajectory tracking control algorithm based on model prediction,and verifies the algorithm through simulation and flight test.The main research contents are as follows:(1)Linear plant protection UAV system test platform was designed.The software and hardware technical scheme was determined by analyzing the operation requirements,and the hardware platform was modular classified.The arithmetic control module composed of the airborne computer and the flight control unit was built,as well as the navigation and positioning module composed of single-line laser,millimeter-wave radar and Sinan integrated navigation system were built.The unmanned aerial vehicle(UAV)ground station is independently developed for real-time monitoring of UAV operation and operation status.A coordinate system is established to accurately describe the motion changes of UAV,and a kinematic model is built based on Newton Euler method.(2)Aiming at problems such as large offset error and unreasonable time allocation of traditional Minimum-snap algorithm,an improved Minimum-snap trajectory optimization algorithm based on linear plant protection UAV kinematics model was proposed.Firstly,the airborne RTK was used to collect the terrain data of the operation plot and build A three-dimensional spatial map.The initial operation path was planned by combining the cow-farming method and A-star algorithm.Secondly,by improving the Minimum-snap algorithm to optimize the job path and combining the initial trajectory state parameters,the time allocation cost function was constructed to obtain the current optimal time allocation.Then,the Minimum-snap trajectory constraint cost function was reconstructed,the position offset gradient penalty factor was added,and the trajectory polynomial coefficients were solved by the optimization method.Finally,the trajectory points are instantiated by combining the position control period and trajectory polynomial to control the motion of the UAV.The experimental results show that compared with the traditional Minimum-snap algorithm,the proposed algorithm can reduce the average acceleration by 7.82% and the average offset error by 45.56%under the same working time.The proposed algorithm has an obvious inhibition effect on the trajectory offset,and reduces the overshoot of the acceleration at the ground steering.The operating trajectory is more accurate and the operating speed is more stable.(3)In order to realize the autonomous tracking operation of linear plant protection UAV,a trajectory tracking controller was designed based on the model predictive control algorithm.In order to improve the trajectory tracking accuracy,the EKF algorithm was used for real-time correction estimation of UAV position.Under the interference of Gaussian noise with external variance of 0.05,the maximum offset of trajectory tracking is 0.09 m,and the tracking effect is good.The algorithm has the characteristics of low hysteresis,high accuracy and no overshoot.In order to meet the real-time obstacle avoidance requirements of UAV,a multi-modal trajectory control scheme is adopted,the control model of the system is reconstructed,obstacle information is introduced into the control model,and the distance between UAV and obstacle is used as a reference index for the optimization of the control system.The autonomous obstacle avoidance tracking controller is designed.The test results show that the maximum effective obstacle avoidance distance reaches 0.55 m without affecting the normal trajectory tracking. |
