Variable-dimensional Flower Pollination obstacle avoidance algorithm on autonomous walking of NAO robot in dynamic environment

Abstract

This paper proposes a novel dynamic obstacle avoidance method based on the Variable-dimensional Flower Pollination (VFP) algorithm, which can avoid dynamic obstacles under an unknown environment. The grid map is used to describe the environment and the shearing map refreshment strategy is used to improve refreshment efficiency. The fitness function is designed by combining Chebyshev distance with Euclidean distance, which can reasonably evaluate the planned path. Simulation results with traditional Flower Pollination algorithm and VFP algorithm are compared. The simulation results show that the VFP algorithm has a faster learning speed than the traditional Flower Pollination algorithm. To verify the simulation result, the VFP algorithm is implemented on the NAO robot, and the experiment result demonstrates that the Variable-dimensional Flower Pollination algorithm is feasible and effective.