How NASA’s Mars Rover is Helping to Chart a Path For Autonomous Trucks
By Dr. Al Kelly
In fact, autonomous vehicle technology has been around for at least 35 years. Research at Carnegie Mellon University (CMU) resulted in an autonomous vehicle successfully crossing the continental United States in 1995, dubbed “No Hands Across America.” Similar efforts have long been underway in Asia and Europe as well.
This research, conducted across the decades by educational institutions such as CMU, Stanford, University of Michigan, and MIT, has been funded by business interests exploring potential AV uses in agriculture and mining, as well as by government agencies such as the U.S. Department of Defense’s Defense Advanced Research Projects Agency (DARPA) for military vehicles like the Family of Medium Tactical Vehicles (FMTVs), and by the National Aeronautics and Space Administration (NASA) for its Mars Rover program.
The key development, going back at least as far as 1987. was lane following, a form of visual servoing technology that enables vehicles to “see” the road ahead and keep them in lane as they travel. Related technologies are stereo vision — depth-sensing technology that uses two cameras just as our two eyes do — and visual odometry, a way to sense the motion of a camera (and the robot attached to it) by watching how things move in the camera image. I was part of a team back in the early 2000s that combined these technologies into an integrated system that could both avoid obstacles and determine vehicle position.
Only a few years later, after much refinement by the broader research community and NASA’s Jet Propulsion Laboratory (JPL), similar algorithms made their way onto the Mars Rovers. Since then, the Mars Rover program has continued to evolve as autonomous technology has improved.
In 1997, the Mars Pathfinder mission deployed the rover Sojourner which went a short distance with human guidance, analyzed planet samples, and took photographs for about three months.
In 2004, the Mars Exploration Rovers A and B, named Spirit and Opportunity, drove around for longer distances directed by human instruction while utilizing visual servoing technology to navigate. Spirit operated successfully for five years, while Opportunity went for more than 14 years.
In 2021, the Perseverance, including its mini helicopter Ingenuity, landed and sent back the best images yet. Its base vehicle design is similar to the Curiosity rover which landed on Mars in 2012. Curiosity has gone much farther and faster and operated more autonomously than its predecessors. As a result of improved algorithms, Perseverance’s software, design, and technology enable the rover to drive itself and determine the best route between points with no direction from Earth.
Many of the scientists and engineers who developed rover technology are today working to help build AVs here on Earth — and some are also taking what they’re learning from developing AVs on Earth and using it to improve rovers for Mars.
There are many reasons why we’re further along in some respects with autonomous driving on Mars than on Earth, starting with safety. The rovers are designed to handle harsh environmental and geographic conditions, but otherwise they operate alone in their stationary environment.
Autonomous trucking engineers, on the other hand, must address a hugely complex, dynamic, high-speed environment that includes human drivers with their errors, lots of other vehicles, traffic lights, and more. It’s all of these additional elements that make it more difficult to build and develop safe and efficient autonomous trucks at scale here on Earth.
While much progress has been made, there is still much to be learned, tested, and verified before fully-autonomous vehicles can be enabled in all applications. Pioneering technologies have opened the door to AVs on Earth and on Mars alike.
At Locomation, we are committed to putting safety above anything else. Our convoying approach maintains a human driver in the leader truck, thereby reducing the “cognitive demand” on the follower truck to make complex judgements by itself. This is why we believe it is the safest and fastest way to scale fully autonomous trucks as we continue to expand their capabilities through technology maturation.