The Research Of Mowing Robot Based On Visual Navigation And Tracking Control

Following the development of lawn industry, Mowing robot has theability to drive automatically, and it can be applied in many fields. Atpresent the methods of mowing are focused on boundary informationwhich is determinated with a sensor when the boundary is encircled by acable. In this study the boundary between lawn and road wasdeterminated by CCD visual sensor, based on vision the Mowing robotwas navigative controted.At first the boundary between lawn and road was distinguished inthis study. The process of monochrome image processing was as follows:median filter, minimal error threshold segmentate, fill in the target image,pick up boundery, straight line fit. The simulation results show that thealgorithm fit navigation control system of the mowing robot.Tracking control is another key technique among the researches ofmowing robot. therefore, it is very important to design a control systemwhich has the following characteristics: small tracking error, quickdynamic response, adaptability to complicated circumstances androbustness and so on. Based on this, personification controller of fuzzycontrol methods is adopted. the simulation results in various complexcondition show that the control system have the competence of pathtracking and resisting random noise.According to mission demand of the mowing robot, a symstem ofpath tracking and boundary recongnition was designed. The examinationresults show that when the start velocity was 0.5m/s, elevation angle was18°, acquisition frequency was 2 Hz, tracking error was allowable withina range.