Skip to content

Latest commit

 

History

History
35 lines (27 loc) · 1.61 KB

driving_curiosity.md

File metadata and controls

35 lines (27 loc) · 1.61 KB

Driving Curiosity: Mars Rover Mobility Trends During the First Seven Years

Authors: Arturo Rankin, Mark Maimone, Jeffrey Biesiadecki, Nikunj Patel, Dan Levine, Olivier Toupet

Year: 2020

Notes:

  • Review of MSL mission (exploring Gale crater and Mount Sharp)
  • Curiosity = 6 wheels rover, that can rotate in both direction independantly (except middle wheels that are not steerable)
  • Visual odometry enables SLIP estimates: evaluate the slip fraction by comparing the motion of VO with the one of wheel odometers
  • Excessive SLIP may raise a fault or even stop the sol
  • 99,55% of VO convergence

Categories of VO failure:

  • Failed to converge (mostly due to a lack of features)
  • Bug in step truncation (FSW is supposed to truncate large step to prevent from VO failure)
  • Sequencing failure: Rover planners may send commands outside of the normal bounds
  • Strategy filure
  • IMU parameter failure

Drives Modes:

  • the whole VO process takes 47s on average
  • Autonomous mode alternates between full VO (ie one iteration every meter) and slip checks (every 20 meters)
  • Visual target tracking ie visual servoing. Tracking is performed with a correlation function cf Overview of the Mars Exploration Rovers’ Autonomous Mobility and Vision Capabilities (already available on MER)
  • Path planning with D* algorithm
  • Details of all drive failure cause (slip limits, time of the day exceeded, obstacle, tilt limit)
  • TRCTL adapt each wheel speed to fit the topography: enables longevity of wheels and actuators

Commentaires:

Belle histoire, considération intéressante sur la mission et les causes d'échec. Mais HS pour la thèse