A Mobile Robot Path Planning Scheme for Dynamic Environments

Abstract

This paper proposes a four-direction search method for obstacle avoidance of mobile robots, and the collision energy of obstacles is modeled based on the neural network. For comparison and discussion purposes, the different invariant step lengths are tested in the static environment. To take the advantages of the large and small step lengths effectively, a variable step length method is adopted to improve the performance, such as less iteration number and lower total energy. The variable step length method is also applied to the dynamic environment to explore the real-time performance for path planning. Simulation results demonstrate the effectiveness and practicability of the presented scheme.