System Design and Obstacle Avoidance Algorithm Research of Vacuum Cleaning Robot

Abstract

Designing multi-sensor hardware system based on Cortex-M0 controller, detecting the surrounding of vacuum cleaning robot by sensor combination of ultrasonic, infrared, collision, etc., we proposed an information fusion algorithm with Kernel PCA based on SFC neural network to process the multi-sensor data. And the output result of SFC neural network is used to control the localization and obstacle avoidance of vacuum cleaning robot. The experimental results demonstrate that the multi-sensor hardware system and obstacle avoidance algorithm based on SFC neural network proposed in this paper can improve the localization and obstacle avoidance accuracy of vacuum cleaning robot highly, which is robust for different working environment of vacuum cleaning robot.