Obstacle Avoidance Algorithm Based on Human Experience Knowledge

This paper studies the obstacle avoidance problem of mobile vehicles. Most of the existing algorithms for obstacle avoidance are proposed for specific environments, lacking adaptability for various environments. Motivated by mimicking the obstacle-avoidance actions of humans in different environments, this paper proposes a human experience knowledge (HEK) based obstacle avoidance algorithm. The proposed algorithm provides reliable support for mobile vehicles to implement tasks in a complex environment. Specifically, we first characterize the related experience knowledge of humans by converting to categorical variables. Then, a logistic regression method is utilized to model the corresponding knowledge based on categorical variables. Finally, the HEK-based algorithm, which dynamically adapts to different situations, is designed according to the corresponding knowledge model. Extensive comparative simulation results are conducted to demonstrate that the proposed algorithm has better adaptability and flexibility for various static obstacle environments and a low computational cost.