Automated maze solving using fluid mechanics based numerical approach

Abstract

Maze solving has direct applications like Robot path finding, emergency evacuation, and drug discovery. Traditional methods of maze solving treat the problem of maze as a directional graph by connecting the key nodes. Because of the directional graph treatment, which is applied on the simplified form of the original maze, the maze solving algorithms at present cannot be directly applied in the critical domains of emergency evacuation etc. The proposed method uses the original complex maze without simplifying and hence can be directly applied in any critical application related to path finding. Proposed approach solves the mazes in a fully automated mode by combining the techniques of image processing, computational geometry and computational fluid dynamics. Moreover the high amount of human intervention needed for generating graphs from mazes in the traditional algorithms is eliminated altogether.