Open Access Open Access  Restricted Access Subscription Access

Path Planning by Using 3-Demensional Map Including Vertical Information

Tomoki Tajiri

Abstract


In this paper, we have tried generating a path from the start position to the goal position which is movable with mobile robots. The shortest path is usually the best route. But there may be high steps or steep slopes in a real environment. It is not possible for the mobile robot with wheels to move in the direction because there is the steep slope or the high step. Q-Learning is one of the method witch is often used to planning path. We have made a robot learn to get path which has less gradient of ground with Q-Learning. The effectiveness of this method is investigated with an actual robot in the real environment that has steps and slopes. The experimental result shows the effectiveness of this approach.

Full Text:

PDF

References


Yoshiaki, S. 2004. Journal of The Robotics Society of Japan. Vol.26. No.4. pp302-305.

Kaelbling, L. P., Littman, M. L., and Moore, A. W.. 1996. Journal of Artificial Intelligence Reserch 4:237-285. Reinforcement Learning: An Introduction. MIT Press.

Thrun, S., Burgard, W. and Fox, D.. 2007. Probabilistic Robotics (Intelligent Robotics and Autonomous Agents series).

Moravec, H. P.. 1988. AI Magazine. Vol.9. 61-74. Sensor Fusion in Certainty Grid for Mobile Robots.




DOI: http://dx.doi.org/10.21535%2FProICIUS.2011.v7.333

Refbacks

  • There are currently no refbacks.