Research On The Complete Coverage Path Planning Of Cleaning Robot

Cleaning robot is one of the reflection of automation control in modern human life, it combines the knowledge of motion control technology, information processing technology and artificial intelligence and other multi-disciplinary knowledge. The key techniques include sensing detection technology, path planning technology, automatic recharge technology, power supply, etc. The path planning technology is a key index to realize intelligent cleaning robot, is also the hot issue of cleaning robot research. Cleaning robot path planning technology requirement is traverse all regions except obstacles in the working area. The paper puts forward a new complete coverage of cleaning robot based on unit decomposition method.In order to realize complete coverage path planning, cleaning robot must has the ability of creating the environment map and self-localization. Environment map is created by edge learning methods, using infrared sensor of robot itself as well as the collision switch access to environment information, the ultrasonic sensors for environment obstacle boundaries of distance information, and then draw molecular area and establish sub-region map by grid method. Cleaning robot self localization through encoder and gyro sensor information. Due to the error of the sensor itself is inevitable, so for cleaning robot system using extended kalman filtering data of senor information fusion, obtain the optimal position estimation in the process of cleaning robot movement, the effectiveness of the algorithm is verified by simulation.Through edge learning, cleaning robot get sub-region, using alternate traversal template for area of coverage, and according to the principle of minimum energy and time to determine the direction of the cleaning robot moving. This paper puts forward the complete coverage path planning algorithm is based on the thought of edge learning– sub-region division – sub-region traversal – edge learning. During the traverse of sub-region, sub-regions formed by isolated obstacles which initially unable to traverse judgment information detected by the sensor, and supplement of its traversal after the preliminary search is terminated. At last, the complete coverage of work area has been done.This paper builds the experiment platform of cleaning robot based on the complete coverage path planning algorithm, including the design of hardware and software. The motion model of cleaning robot consists of two-wheel differential drive system with a wheel encoder for each wheel, and another caster help to support the body and turn around. Control system including power supply module, serial communication module, motor driver module, sensor information collection module, etc. Software design including motor drive PWM waveform, the sensor information acquisition and complete coverage path planning algorithm, etc.