| As one of the key technologies of agricultural production,variable spraying technology can provide technical guarantee for fertilizer and pesticide reduction and efficiency.However,the variable spraying technology of agricultural UAVs is not fully mature yet,and the precise regulation of UAVs for variable application of chemical solutions based on prescription maps is a current research hotspot in this field.In this paper,with the goal of realizing the variable spraying technology based on agricultural UAVs,through embedded control technology and neural network PID flow regulation method to study the three aspects involved in the process of agricultural UAV variable spraying of liquid fertilizers and pesticides: variable prescription map processing,design and construction of agricultural UAV variable spraying system,and variable spraying experiment and evaluation.The main results include:(1)A prescription map interpretation scheme was designed for the variable spraying operations of agricultural UAV.The variable prescription map was written into the NAND flash inside the ARM chip,and the GPS coordinate forward pre-processing was performed by coordinate system rotation transformation.In the process of operation,the network RTK(RealTime Kinematic)technology was used to obtain the centimeter-level GPS positioning information of agricultural UAV,and then combined with the raster recognition algorithm for addressing matching,and set the target spraying volume of the system by calling up the prescription decision information in the prescription map according to the real-time position of the UAV.(2)The parameter optimization of the incremental PID control algorithm was realized by using the RBF(Radial Basis Function)neural network,and the RBF-PID,incremental PID and single neuron PID control algorithms were simulated and compared through the simulation platform.The algorithm simulation comparison results showed that the RBF-PID control algorithm had the fastest adjustment speed in the step response simulation,the rise time was only 0.007 s,and there was no overshoot and static error;in the square wave tracking simulation,the average rise time of the RBF-PID control algorithm was 0.044 s,The tracking error was5.47%,and the tracking speed continued to accelerate with the increase of the number of adjustments.It had a good self-learning ability and is suitable for agricultural UAVs to perform variable spraying operations in the state of high-speed movement in the air.(3)Through the development of hardware system and software system,an agricultural UAV variable spraying system was developed.The system selected STM32H7 series highperformance single-chip microcomputer as the control center of the variable spraying system,used the incremental PID controller to form the closed-loop control loop of the spraying amount,and controled the PID according to the deviation between the actual spraying amount and the target spraying amount through the RBF neural network.The parameters of the chemical pump were optimized and adjusted,and the PID controller controled the PWM(Pulse Width Modulation)signal with continuously adjustable duty cycle of the microcontroller,which further controled the input voltage value of the liquid pump to changed its rotational speed and realize the precise variable adjustment of the spraying volume.(4)The developed variable spraying system was tested in the indoor spraying volume following experiment and field variable spraying experiment.The indoor experiment results showed that the spraying change process was fast and stable with a maximum overshoot of3.88%,an average rise time of 0.24 s,and a maximum average absolute error of 5.85% when the variable control was controlled by the RBF-PID control algorithm;the field experiment results showed that the system had an average overshoot of 3.21%,an average rise time of 0.47 s,and a variable control error of 1.36%,there was a good spraying effect when the spraying volume is set above 0.7L/min,and the three parameters of droplet deposition,coverage rate and coverage density all showed that the system can correctly perform the variable spraying task according to the prescription map settings.The results of the study can provide a basis and reference for the improvement of variable spraying technology for agricultural UAVs. |
