Robot motion has been considered as a solved problem in D* algorithm implementations but it has important effects in efficiency of algorithm for generating shortest path with least displacement to reach goal in action. This paper presents Optimized D* algorithm with Two phase motion in unknown damaged indoor environments. Its aim is to find shortest path to goal position through long corridors, tiny doorways and wide and complicated obstacles. Most implementation considered environment in grids which will cause linear paths that are not smooth and collision with sharp edged obstacles. We used a semi-grid model for representing areas around robot. In our model, both features of discrete and concrete environments are used. According to our model, there are two phases, first one is turning phase and second one is moving phase. By these two motions, we optimized motion of robot in D* algorithm. The research is divided into two stages. The first stage is implementation in 2D grid-based and the second stage, in 3D simulator with P3AT robot. Results obtained from various aspects, have proved robustness and accuracy of our work.
展开▼
