Improved Rover State Estimation in Challenging Terrain

  • Authors:

  • Brian D. Hoffman;Eric T. Baumgartner;Terry L. Huntsberger;Paul S. Schenker

  • Affiliations:

  • Silicon Graphics Computer Systems, Mountain View, CA 94043;Science and Technology Development Section, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109;Science and Technology Development Section, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109;Science and Technology Development Section, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109

  • Venue:
  • Autonomous Robots
  • Year:
  • 1999

Quantified Score

Hi-index 0.01

Visualization

Abstract

Given ambitious mission objectives and long delay timesbetween command-uplink/data-downlink sessions, increased autonomy isrequired for planetary rovers. Specifically, NASA‘s planned 2003and 2005 Mars rover missions must incorporate increased autonomy iftheir desired mission goals are to be realized. Increased autonomy, including autonomous path planning andnavigation to user designated goals, relies on good quality estimatesof the rover‘s state, e.g., its position and orientation relative tosome initial reference frame. The challenging terrain over which therover will necessarily traverse tends to seriously degrade adead-reckoned state estimate, given severe wheel slip and/orinteraction with obstacles. In this paper, we present the implementation of a complete rover navigationsystem. First, the system is able to adaptively construct semi-sparseterrain maps based on the current ground texture and distances topossible nearby obstacles. Second, the rover is able to matchsuccessively constructed terrain maps to obtain a vision-based stateestimate which can then be fused with wheel odometry toobtain a much improved state estimate. Finally the rover makes use ofthis state estimate to perform autonomous real-time path planning andnavigation to user designated goals. Reactive obstacle avoidanceis also implemented for roaming in an environment in the absence of auser designated goal. The system is demonstrated in soft soil and relatively dense rock fields, achieving state estimates that are significantly improved with respect todead reckoning alone (e.g., 0.38 m mean absolute error vs. 1.34 m), andsuccessfully navigating in multiple trials to user designated goals.