Almost 30 years have passed since the Viking missions of the mid 1970s, the last successful science-driven NASA mission to land on Mars. Indeed, NASA is hoping for a large scientific return with the 2003 Mars Exploration Rover (MER) mission that is under way.
The six-wheeled twin rovers are serving as trusty robotic field geologists, scouting for evidence the red planet once was or is a habitable place. To search for signs of past or present liquid water -- a medium for life as we know it on Earth -- the rovers are equipped with nine cameras and an instrument-packed robotic arm similar in size to a human -- to measure the mineral makeup of martian rocks and the make-up of the atmosphere.
While the rovers will operate as human hands and eyes on Mars, the robots are dependent almost entirely on human intelligence on Earth.
During surface operations scheduled to start in January, the MER rovers will be under the constant supervision and care of 240 scientists and engineers. The staff will work around the clock to analyze the data the rovers return to Earth, to choose science goals for the rovers and to build plans that tell the rovers what to do and how to do it on the next martian day.
NASA is working to develop software that will give rovers more autonomy so that missions can conduct more science.
MAPGEN (Mixed Initiative Activity Planning Generator) is a ground-based decision support system for MER mission operations and science teams that begins to give content to the notion of autonomous planetary exploration.
MAPGEN is called a mixed initiative planner. The essence is in human interaction with the advanced planning software. While the routine plan generation process is handled by the machine, the user brings unique knowledge and experience to bear to produce qualitatively good plans, said MAPGEN project lead Kanna Rajan.
The paradigm is to enable the person using the software to critique a plan that the system automatically produces and ensure that resulting plans are viable within mission and flight rules.
MAPGEN will determine that the plan is within the bounds of the available resources onboard the rovers, and is free of any temporal violations. In addition to taking into account various mission and flight rules, the system keeps track of the intent of the user when generating a feasible plan.
If, for example, MAPGEN finds that executing the plan would drain the rover's battery midway, the software rearranges the tasks in order to remove the violation, generating a feasible plan. Another software downstream of the use of MAPGEN, RSVP, with the help of mission operations teams will build the actual sequences that will be uplinked to the rover.
Does MAPGEN sound like a useful tool? Its developers persuaded NASA managers to accept MAPGEN into the mission. That's because of the high cost of missions -- MER is estimated at $800 million -- and the risks associated with sending rovers far away to traverse uncertain terrain. Mission managers approach new technologies with caution.
"We spent eight months proving ourselves," Rajan said. "We really had to convince the operations and science community because they didn't buy into the utility of what we were proposing. We had to prove it."
MAPGEN uses the planner from a software system that in 1999 demonstrated for the first time that a spacecraft could in effect fly itself. Rajan was a principal member of the team that developed the software, the Remote Agent, which enabled the spacecraft Deep Space 1 to generate a mission plan and execute it onboard with no human supervision. Remote Agent was risky but successful. It opened up the notion that autonomy is useful and it can be done onboard.
Said Bob Morris, NASA's project manager for Intelligent Systems, "MAPGEN is a step in the right direction."