Obstacle Avoidance Algorithm for Indoor Mobility Robot Based on Improved Support Vector Machine
Jie Hao · 2019
With the development of artificial intelligence, robot will be applied to every aspect of our life. In the domain of robotics, especially indoor robotics, the ability of obstacle avoidance is one of the most important questions. In this paper, first of all, some classical machine learning algorithms are applied for indoor robot obstacle avoidance which is in the unknown environment based on multiple ultrasonic sensors. During the move, obstacles will be detected by sensors on the robot, through the machine learning algorithm, the robot will know how to do when the sensor detect an obstacle. Then, proposes an improved Support Vector Machine(SVM) algorithm based on the Grid Search algorithm. This algorithm can accurately help robot find the trajectory for moving. Experimental results also presented to show the algorithm improved the generalization ability and real-time performance