Dynamic map for obstacle avoidance

Abstract

Most of the existing obstacle avoidance algorithms work well in a relatively static environment, which consists of mostly stationary obstacles and maybe some slow-moving obstacles. In a dynamic environment, however, these obstacle avoidance algorithms may result in a robot colliding with a moving obstacle. In this paper, we propose an idea of constructing a dynamic map of obstacles that shows the predicted positions of the moving obstacles in future. It imitates the human actions of monitoring and estimating the movements of other objects in order to choose a suitable path for themselves to travel in a dynamic environment. We used this idea on an existing obstacle avoidance algorithm, the vector field histogram approach. The simulation results show that the resulting method is able to navigate a robot safely in a dynamic environment.