Research On The Control System Of An Intelligent Sprayer For Precision Weed Control In Turf

With the rapid development of the social economy,the level of urban afforestation is also constantly improving,along with the growing planting area of lawns.Due to the relatively wide lawn planting area,the current maintenance of the lawn is relied on spraying herbicides on a large area by manually driving a tractor.This weeding approach is convenient and efficient.However,it wastes herbicides and pollutes the air.It is challenging to use herbicides to control a variety of weeds.Therefore,there is an urgent need for intelligent weeding equipment that can precisely apply pesticides to a single weed in turf.In order to save manpower,remote control driving is selected instead of manual driving,as the sprayer works outdoors.In addition,it is time-consuming to carry out manual remote control,which decrease the work efficiency.The sprayer can work autonomously in turf,which can save manpower and improve weeding efficiency.Therefore,this paper combines the intelligent sprayer with SLAM technology to enable it to have the ability to move autonomously and improve the intelligence of the sprayer.The main research contents are as follows:(1)Design of the remote-control system for the sprayer.On the basis of prototype manufacturing,a two-level distributed control scheme for the handheld remote-control terminal and the vehicle-mounted sprayer was developed.A wireless local area network and an Open Wrt system were built,compiled,and configured in the WiFi module.Afterwards,the network-to-serial port ser2 net application was installed with a programed self-starting script.a client based on the Android system was developed,and a Graphical User Interface was built and communicated via communication protocol.In addition,the corresponding functions of the vehicle-mounted driving system of the sprayer was implemented and integrated into the spray system.The camera acquisition program was designed and deployed the neural network module on the edge AI device Jetson Xavier NX.Finally,the control of solenoid valve was achieved thorough a corresponding communication protocol.The remote-control driving function of the sprayer was verified in a field driving test.Every driving function of the sprayer can operate normally and accurately according to the control instructions issued by the APP.In the driving test,the speed of the sprayer was 0.201m/s.In the intermittent driving mode,the driving distance of the sprayer is in the range of 93-115 mm,which is basically in line with the spraying range of the sprayer nozzle,and the distance error deviation is 8.6%.Two herbicide application experiments were carried out through the actively grown period and the dormant period of the lawn.Water spraying,broadcast spraying,and precision spraying were performed with three weeks of photographing and observation to compare their weed control effect.While there was no significant difference in the rate,precision spraying significantly reduced the amount of herbicide solution used by 78%.(2)Study on autonomous movement of the sprayer.The overall scheme of the autonomous movement control system of the sprayer was formulated.According to the four-wheel structure of the sprayer,a two-wheel differential kinematics model was established,and the trajectory of the sprayer and the 2D laser radar observation model was calculated.The Gmapping algorithm based on RBPF and the Cartographer algorithm based on graph optimization were studied on the basis of SLAM algorithm,and the adaptive Monte Carlo positioning algorithm was studied for sprayer positioning and path planning,and the A* algorithm was employed for global path planning,while for local path planning,the TEB algorithm was utilized.Finally,the software design of the autonomous mobile control system of the sprayer is carried out,and the chassis drive system program is written according to the communication protocol.The self-starting script of the mobile program.The feasibility and accuracy of the algorithm were verified by simulation.The mapping effect and accuracy of the two SLAM algorithms were compared in the autonomous movement test,the results showed that the relative error of the Cartographer algorithm was between 0.44%and 1.20%,which was higher than the Gmapping algorithm,in terms of mapping accuracy and robustness.In the autonomous movement and obstacle avoidance test,the sprayer can avoid obstacles and reach the set target point,the position deviation distribution is within 9.75%,the absolute deviation of the driving heading angle is within 13.76°,and the success rate of autonomous movement and obstacle avoidance is about 85%.The accuracy of AMCL positioning algorithm and path planning algorithm is verified,and the feasibility of using SLAM technology in combination with lawn weeding equipment is verified.