Dynamic Path Planning for Mobile Robots with Cellular Learning Automata

Abstract

In this paper we propose a new approach to path planning for mobile robots with cellular automata and cellular learning automata. We divide the planning into two stages. In the first stage, global path planning is performed by cellular automata from an initial position to a goal position. In this stage, the minimum distance is computed. To compute the path, we use a particular two-dimensional cellular automata rule. The process of computation is performed using simple arithmetic operations, hence it can be done efficiently. In the second stage, local planning is used to update the global path. This stage is required to adapt to changes in a dynamic environment. This planning is implemented using cellular learning automata to optimize performance by collecting information from the environment. This approach yields a path that stays near to the obstacles and therefore the total time and distance to the goal can be optimized.