关键词: 农用无人机, 变量喷雾, 单神经元, PID

Abstract:  An agricultural UAV variable spraying system based on singleneuron PID control is designed for the problems of long response time, large overshoot, and following unstable effect of the existing agricultural UAV variable spraying system. The system adopts the sensor detection flow information as the control basis and uses the single neuron selflearning ability to continuously adjust the PID parameters to precisely regulate the spray flow and achieve the goal of fast and stable variable regulation. In order to verify the actual variable control effect of this system control algorithm, Matlab was used to simulate and analyze the traditional PID and singleneuron PID control algorithms. During the indoor spraying experiment, the continuous variable effect of singleneuron PID and traditional PID was compared. During the field test, the actual variable spraying effect was detected using flow sensors, and watersensitive paper was used to obtain the droplet deposition distribution. The algorithm simulation results show that the singleneuron PID control algorithm is better than the conventional PID in terms of rising time, with a shorter delay time and no steadystate error. The indoor spraying experimental results show that the spraying change process is stable and rapid when using singleneuron PID for variable control, with a maximum overshoot of 3.88% in spray flow rate, an average rise time of 0.85 s, and an average absolute error of 7.13%. The field variable spraying test results show that the average rise regulation time of the system is 0.87 s, and the spraying volume control error is within 5%. The droplet test results show that the deposition of droplets increased with the increase of spray flow rate. The results of the study can provide a theoretical basis for the development of variable spraying technology.

Key words: UAV, variable spray, single neuron, PID