Uniform and Efficient Exploration of State Space Using Kinodynamic Sampling-Based Planners
Sampling based algorithms such as RRTs have laid down the foundation for solving motion planning queries for systems with high number of degrees of freedom and complex constraints. However, lack of balanced state-space exploration of RRTs calls for further improvement of these algorithms. Factors such as drift, underactuation, system dynamics and constraints, and the lack of an energy/time based distance metric in state space can cause RRT propagation to be uneven. This paper focuses on improving the coverage of the RRT algorithm for physical systems that demonstrate a tendency to restrict the growth of RRT to certain regions of the state space. A localized principal component analysis based approach is proposed to learn the propagation bias of state-space points sampled on a grid during an offline learning phase. To compensate for this bias, expansion of the RRT in real-time is steered in the direction of the least principal component of the propagation of the state-space sample selected for expansion. The algorithm is tested on various systems with high degres of freedom and experimental results indicate improved and uniform state-space coverage.
KeywordsUniform state-space coverage Local-PCA based RRT
The authors would like to thank their colleagues Yanbo Li and Kostas Bekris for providing guidance and assisting with replicting results from .
- 3.Choset, H.M.: Principles of robot motion: theory, algorithms, and implementation Intelligent robotics and autonomous agents. MIT Press, Cambridge (2005)Google Scholar
- 5.Glassman, E.L.: A quadratic regulator-based heuristic for rapidly exploring state space. Master’s thesis, Massachusetts Institute of Technology, Massachusetts (2010)Google Scholar
- 6.Hsu, D., Kindel, R., Latombe, J.C.: Rock . Randomized kinodynamic motion planning with moving obstacles (2000)Google Scholar
- 7.Jolliffe, I.: Principal component analysis, 2nd edn. Springer, New York (2010)Google Scholar
- 8.Ladd, A.M., Kavraki, L.E.: Motion planning in the presence of drift, underactuation and discrete system changes. In: Robotics: science and systems I, pp. 233–241. MIT Press, Boston (2005)Google Scholar
- 9.LaValle, S., Kuffner, J.: Rapidly exploring random trees: progress and prospects. In: Proceedings Workshop on Algorithmic Foundations of Robotics WAFR, pp. 293–308 (2001)Google Scholar
- 10.LaValle, S.M.: Planning Algorithms. Cambridge University Press, Cambridge (2006)Google Scholar
- 11.Lewis, F.L.: Applied optimal control and estimation: digital design and implementation. Prentice Hall, New Jersey (1992)Google Scholar
- 12.Li, Y., Bekris, K.E.: Balancing state-space coverage in planning with dynamics. In: IEEE International Conference on Robotics and Automatio (ICRA), pp. 3246–3253 (2010)Google Scholar