Rover Obstacle Visualizer and Navigability Evaluator (ROVANE)
The Situation
Planetary exploration using rovers—such as NASA's rovers currently on Mars—calls for scientists in their labs on earth to specify new rover destinations on a daily basis. Mission scientists-whose studies and experiments are carried out using the rovers-are not experts in rover navigation and depend on expert advice from the ground operations team at NASA-JPL (Jet Propulsion Laboratory). The turnaround time on human advice would make this a very inefficient process. Scientists would be forced to make overly conservative choices because of insufficient time to evaluate all the interesting paths.
The Charles River Analytics Solution
We developed a software tool (ROVANE) that allows scientists to rapidly evaluate navigability for themselves, before submitting a destination for final verification by NASA's engineering team. We demonstrated the concept by extending NASA-JPL's Web Interface for TeleScience (WITS) to provide immediate visual feedback on the amount of maneuvering required to reach any destination. ROVANE generates an obstacle map (see figure 1) from stereo imagery captured by the rover. Using this map, ROVANE then computes on demand the shortest obstacle-free path to user-selected destinations (figure 2). ROVANE's robustness stems from our obstacle detection algorithms that fuse evidence from both color cues in the images and 3D shape information derived from stereo.
The Benefit
ROVANE was extensively tested in three annual field tests where NASA-JPL dropped a rover somewhere in the Mojave Desert and had a team of scientists run Mars-like missions in unknown terrain. The scientists used WITS for all data analysis and rover commanding. The ROVANE component got high praise from all users. WITS and the feedback from these field tests was the basis for the software tool currently in use by NASA scientists to design goals for the Mars Exploration Rovers and analyze the images received from Mars. A public version of this tool, Maestro, is available for download from NASA-JPL.
Related Papers
Goodsell, T., Snorrason, M., Ruda, H., and Ablavsky, V. (2001) "Navigability Evaluation and Visualization for Mars Rover Operations," Proceedings of the International Conference on Visualization, Imaging and Image Processing, Marbella, Spain (July). Abstract
Goodsell, T., Snorrason, M., Ruda, H., and Ablavsky, V. (2001) "Automated Obstacle Mapping and Navigability Analysis for Rover Mission Planning," Proceedings of SPIE, Volume 4364, AeroSense, Orlando, FL (April). Abstract
Goodsell, T.G., Snorrason, M., Ruda, H., and Ablavsky V. (2000) "Rover Obstacle Visualizer and Navigability Evaluator," Proceedings of SPIE Vol. 4195, Photonics East, Boston, MA (November). Abstract
Snorrason, M., Norris, J., and Backes, P. (1999) "Vision-based Obstacle Detection and Path Planning for Planetary Rovers". Proceedings of SPIE, Vol. 3693, AeroSense, Orlando, FL (April). Abstract
