Development Of A Multifunctional And Autonomous Agricultural Robot

In the mid-1990s, China began the research and development of agricultural robotics and has made significant achievements so far. However, Agricultural robots are difficult to popularize in agricultural production because of the special nature of the agricultural envi-ronment and the limitations of technology. Even in the developed countries agricultural ro-bots also cannot achieve a truly effective and universal use. One reason is that many of the key technologies need further improvement and innovation. The other reason is that the ef-ficiency of agricultural robots is very low and the maintenance cost is high due to the sea-sonal. With the gradual intensification of China’s aging population, the number of popula-tion engaged in agricultural labor will become less and less and labor costs will be higher and higher. Therefore, it is increasingly important to achieve the popularity and practicality of agricultural robots. One of the effective ways to solve these problems above is to de-velop a multifunctional and autonomous agricultural robot to apply to agricultural produc-tion. So research on multifunctional agricultural robots with universal applicability, scal-ability and low-cost has great significance for agricultural robots’ practical popularity.This paper developed a multifunctional and autonomous agricultural robot for field operations with open ideas. The robot’s autonomous walking is guided with machine vision. The robot is mainly for leafy vegetables’ with line characteristics spraying and weeding. This paper is mainly based on the following two points to develop the multifunctional au-tonomous agricultural robot:(1) The robot’s operations environment and operations tasks to design.(2) The idea of open architecture to design.The main contents including:1. The overall program design of the multifunctional and autonomous agricultural robot is mainly including the reasonable consideration and layout of implementing bodies, control systems, navigation systems and the implementation of devices. At last, the agri-cultural robot-driven program, technology parameters, navigation and system control method were determined.2. The operational tools and execution device were designed and manufactured, which is mainly for crop spraying and mechanical weeding operations. The execution de- vice is easy to replace to meet different operational requirements. The operational tools and the execution device model were established utilizing the AutoCAD and Pro/e software.3. The control system program was designed. The hardware and software of control system were both described. The microcontroller-BasicATOM Pro40m was chosen based on open characteristics of the control system and also a detailed introduction was made. Finally, the flowchart of software architecture of autonomous walking and spray operations were designed. The coupling circuit and the like were also designed to solve problems for the DC motor drive.4. Machine vision system. Technologies on machine vision navigation and identifica-tion were studied. In this paper, the machine vision was selected to guide the robot as well as crop and weed identification. The Hough transform and vertical projection method were used for image processing to access to the center line of the crop rows. Finally, the method was determined after analysis and comparison. The weeds and crops were identified by morphological characteristics. In this paper, all images were processed with the MAT-LAB(2012a) software.5. Test and results analysis. Two tests were made on the developed multifunctional and autonomous agricultural robot mainly including autonomous walking ability and op-erations effectiveness. The simulation test and field trials were done respectively. Before the test, the camera and agricultural robot’s calibration were done and the calibration data were analyzed and integrated. Finally, the test results were analyzed and improvements were suggested.The test results show that the autonomous agricultural robot can basically achieve the desired functionalities. When the agricultural robot’s initial speed is0.2m/s, the average de-viation of the navigation is about18.8mm in Simulations test and38.5mm in field test. The ef-fect of spray was also well. The test results were also analyzed and summarized, based on which further improvement and optimization recommendations were made.