Swarm Intelligence Based Maze Solver

Abstract

Efficient collision-free motion through environments comprising obstacles is crucial to developing autonomous navigation systems used for target detection and tracking. In this paper, we have designed a two-robot system with swarm intelligence, whereby one bot helps the other autonomously navigate through a maze and arrive at the destination. The bot commencing the navigation uses infrared sensors to avoid obstacles whilst following the Left Wall Follower Algorithm. Upon reaching the target position, the initial bot wirelessly transfers the unoptimized path information recorded while traversing the Maze to the second bot, which then extracts the optimum path and autonomously follows it. We have used short-range Bluetooth for communication and Arduino Uno chip-set to program both the bots. Rather than stepper motors, we have opted for direct current motors together with differential kinematics. This hardware modification avoids slippage, reduces cost, and increases the bots' degrees of freedom and speed. Furthermore, establishing a communication pathway between two Arduinos using a Bluetooth module for coordinated search with a novel encoding algorithm is a significant milestone of this project.