The Research On Intelligent Control Technologies For Automatic Robot Mower

This paper mainly considered on the automatic intelligent robot mower (IRM), and focusedon the lawn recognition of the robot mower, non-contact avoidance, automatic returning to thecharge algorithms and other aspects of intelligent automatic control technologies. After a briefintroduction to the studies and applications of current intelligent robot mowers, a control solutionfor intelligent robot mower based on computer vision technology was suggested in this paper.Lawn recognition technology is a major breakthrough and innovation of the intelligence levelof intelligent robot mower. This Paper combined the image recognition and the color edgedetection algorithms, and used the ratio of the intrinsic property of the grass color and edge pixelsto implement the lawn recognition based on statistical analysis, using the single camera on themowing robot. Then, a steering angle feedback for the closed-loop control with the feature pointtracking algorithm based on SURF was proposed, to solve problems of the traditional open-loopcontrol of the drive motors and wheels slippery. A set of ultrasonic obstacle avoidance hardwarewas designed. And combining with grass identification, we proposed an intelligent non-contactavoidance algorithms, decelerating when obstacle is close and steering to avoid obstacles and non-grass areas. Due to the current intelligent robot mowers depend on the boundary line or manuallysetting the virtual boundary, we used grass recognition technology while could enable borderlesswork. Also using the image pattern recognition technology, we’ve designed the charging base, itsrecognition algorithm and several intelligent vision-based automated search programs: the S typesearch, the sector type search, the diffusion type search, as well as the along the border search. Theintelligent robot mower used the above programs to capture and track targets while walking. Whenapproaching the charging base, the robot could get its position and the angle information by thealgorithms. The IRM would adjust its position to complete the automatic docking.As the special nature of this research, the paper made a number of algorithms based oncomputer vision, image processing and intelligent control technologies for the intelligent robotmower. After lots of testing for the above algorithms, it proved a certain value for the control technologies of intelligent robot mower.