Practical Design of a Path Following for a Non-holonomic Mobile Robot Based on a Decentralized Fuzzy Logic Controller and Multiple Cameras

This paper proposes an implementation of a practical model-free path planning and path following algorithm for a non-holonomic indoor wheeled mobile robot using multiple cameras and a decentralized fuzzy logic controller. The proposed algorithm is divided into three stages. The first stage uses the multi-stencils fast marching (MSFM) path planning method. In general, the path resulted from the direct implementation of fast marching methods does not guarantee to be safe or smooth. Subsequently, the robot can touch corners, walls and other obstacles. The proposed algorithm adds a preprocessing stage before the MSFM planning method, based on robot dimensions in order to solve such problem. The second stage uses the visual information extracted from the images captured by multiple cameras, in order to estimate the position and orientation of the mobile robot at each frame. The third stage uses a decentralized control strategy, with three identical proportional derivative-like fuzzy logic controllers (PD-like FLC) connected in parallel, as a path follower to keep up the robot on the desired path. The obtained experimental results in this paper show that the developed design is capable of estimating the shortest path efficiently, while avoiding obstacles and guiding the robot to follow the path in real time.