Current planetary rover localization techniques are lacking in autonomy and accuracy. An autonomous method of globally localizing a rover is proposed by matching features extractedvfrom a 3D orbital elevation map and rover-based 3D lidar scans. Localization can be further improved by including odometry measurements as well as orientation measurements from an
inclinometer and sun sensor. The methodology was tested with real data from a Mars-Moon
analogue site on Devon Island, Nunavut. By tying 23 real scans together with simulated odometry over a 10km traverse, the algorithm was able to localize with varying degrees of accuracy.
Output uncertainties were large due to large input uncertainties, but these could be reduced in future experimentation by minimizing the use of simulated input data. It was concluded that the architecture could be used to accurately and autonomously localize a rover over long-range traverses.
| Identifer | oai:union.ndltd.org:TORONTO/oai:tspace.library.utoronto.ca:1807/17497 |
| Date | 30 July 2009 |
| Creators | Carle, Patrick J. F. |
| Contributors | Barfoot, Timothy D. |
| Source Sets | University of Toronto |
| Language | en_ca |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.002 seconds