A Dynamic Model for Assessing Rover Mobility
- Paper ID
5767
- author
- company
Washington University in St. Louis
- country
United States
- year
2010
- abstract
Since April 2009 until this writing (December 2009) the Mars Exploration Rover (MER) Spirit has been embedded in sandy material on the west of Home Plate at a site called “Troy”. While traveling south toward Goddard and Von Braun for the winter, Spirit skirted the edge of the 8 m wide Scamander crater and broke through a thin surface crust and became embedded in loose sulfate-rich sands. Four of Spirit’s six wheels became partially-to-nearly-completely buried, percent slip was on the order of 95-98\%, and images of the underbelly showed that the warm electronics box may be high-centered on a rock. Ground-based testing was conducted until October 2009 using engineering versions of Spirit at JPL in a mobility testing laboratory. Concurrent was the start of a modeling effort to better understand vehicle dynamics and to aid extrication efforts. MSC Adams is a commercial soft-ware package that can dynamically model mechanical systems and their interactions with a phys-ical environment. The baseline Adams model uses contact forces with Coulomb frictional con-tacts and Stribeck stiction and friction transition velocities, with vertical deformation modeled as a simple damped spring system. Each contact between the mechanical system (i.e. rover wheel) and its environment (i.e. various terrain elements) is defined and incorporated into the overall dynamic model. We are currently correlating Adams model results with ISIL test data and replicating the drives into Troy using simulated elevation maps generated from flight data. The simulations replicate yaw about the inoperable right-front wheel, translational movement in the down-slope direction, excessive (about 90\%) slip in the drive direction, and a ten-dency to “pop a wheelie” with the right middle wheel. Low actuator currents associated with recent extrication events for Spirit suggest a similar process has occurred, but without a full wheelie. We are also developing a more realistic deformable soil model and topographically based scenes with strewn rock fields. Rover motions will also leave a record? of the deformed soil properties and of any new topographic elements generated by the drives. It is anticipated that this model will be used in a predictive manner to assess terrain navigability and will become part of the overall effort in path planning and navigation for both Martian and lunar rovers.