A new active wave computing based real time mobile robot navigation algorithm for dynamic environment

Abstract

This paper introduces an active wave computing based algorithm for real-time robot navigation problem in dynamically changing environment and presents a real test setup with a roving robot. Travelling waves propagated on Cellular Neural Network (CNN) were employed by the system to find the shortest path as proposed in our previous work. The setup has a ceiling camera which captures an image of the top view of the platform at almost every one second. After preprocessing of this image, it is loaded to the network and an active wave which is generated by this network is propagated until the robot is covered by the wave. This method reduces the simulation time and speeds up the system. Although the robot is navigated in a dynamically changing environment, the system does not need to check the moving obstacles. The algorithm proposed here always generates a solution for the current state of the platform without the information about its past states. At every repetition of the algorithm, the target position is updated by the system if it is changed. Any change of the target position is immediately applied to the location of the wave source. Hence the system becomes working in dynamic environment with dynamic target.