Weeding Control System for Variable Target Spraying Based on Fuzzy Control

Abstract

Abstract A variable target weed spraying system using fuzzy control algorithm which based on processor S3C2410 was designed in order to reduce herbicide dosage in cotton fields. This system achieved automatic weeds identification, system speed and flow information collection, flow and pressure adjustment and other functions. According to spraying machine speed and weed size factors that affect variable spraying, a mathematical model of variable spraying factors was established. A dual-input single-output fuzzy controller was designed and simulated by Matlab software. An indoor variable spraying test rig was designed, combined with the test data about the relationship among volume, speed and weed size. A double input single output controller based on Adaptive Neural Fuzzy Inference System (ANFIS) was designed. The controller related hardware circuit and program flow were also designed. Simulation results showed that the controller could control the amount of spraying with the speed and density in real-time changed. The controller was simulated according to the test data, an indoor test-bed which could simulate cotton environment was built and the spraying system was tested. The results showed that the opening time of nozzle was 0.2s, the spraying machine speed was between 0~1m/s, the weed size was between 0~100cm2, the controller changed the spraying amount from 0~4mL.