NASA engineers are planning to add a strong dose of artificial intelligence
(AI) to planetary landers and rovers to make these robot
spacecraft much more self-reliant and capable of making basic decisions
during a mission without human control or supervision. In the past, robot
rovers contained very simple AI systems, which allowed them to make a
limited number of basic, noncomplicated decisions. In the future, however,
mobile robots will possess much higher levels of AI or machine intelligence
and be able to make decisions now being made by human mission
controllers on Earth.
One of the technical challenges that robot engineers face is how
to encapsulate the process by which human beings make decisions in
response to changes in their surroundings into a robot rover or complex
lander spacecraft sitting on a planet millions of miles (kilometers) away.
To make the detailed exploration of the Moon and Mars by mobile robots
practical over the next two decades, future robot rovers will have to be
intelligent enough to navigate the surface of the Moon or Mars without a
continuous stream of detailed instructions from, and decision-making by,
scientists on Earth.
Large teams of human beings on Earth are needed to direct the Mars
Exploration Rovers (MER) Spirit and
roll across the terrain of Mars looking for evidence of water. In a very slow and deliberate process, it takes human-robot teams on
two worlds millions of miles (kilometers) apart several days to achieve
each of many individual mission milestones and objectives. Specifically,
it takes about three (Earth) days for the Spirit or
to visualize a nearby target, get to the target, and do some contact science.
Mars in terms of travel up to 330 feet (100 m) per Martian day (sol) across
the surface of the planet. (A sol is a Martian day and is about 24 hours, 37
minutes, 23 seconds in duration, using Earth-based time units.) Imagine
trying to explore an entire continent here on Earth using a system that
travels a maximum distance each day equivalent to the length of just one
football or soccer field.
This chapter examines how in future a mobile robot with more
onboard machine intelligence (or AI) will collect data about its environment,
and then make an on-the-spot evaluation of appropriate tasks
and actions without being dependent upon decisions made by humans.
Advanced AI systems on board such smart future mobile robots will eventually
allow them to mimic human thought processes and perform tasks
that a human explorer would do. For example, such smart rovers might
pause to make an on-the-spot soil analysis of an interesting sample, communicate
with an orbiting robot spacecraft for additional data about the
immediate location, or even signal other robot rovers to gather (swarm)
at the location in order to perform a collective evaluation of the unusual
discovery.
Within the next two decades, teams of smart robots, interacting with
each other, should be able to map and evaluate large tracts on the surface
of the Moon or Mars. An interactive team of smart robot rovers would
provide much better coverage of a large area of land and perhaps even
exhibit a level of collective intelligence while performing tasks too difficult
or complex for a single robot system. With a team of robots, the mission
objectives can be accomplished, even if one robot fails to perform or is
severely damaged in an accident.
Prospecting for Lunar Water
The Moon is nearby and accessible, so it is a great place to try out many
of the new space technologies, including advanced robot spacecraft,
which will prove critical in the detailed scientific study and eventual
human exploration of more distant alien worlds, such as Mars. Whether
a permanent lunar base turns out to be feasible depends on the issue of
logistics, especially the availability of water in the form of water ice. The
logistics problem is quite simple. Water is dense and rather heavy, so shipping
large amounts of water from Earth’s surface to sustain a permanent
human presence on the Moon this century could be prohibitively expensive.
Establishing a permanent human base on the Moon becomes much
easier and far more practical if large amounts of water (frozen in water ice
deposits) are already there.
This unusual resource condition is possible, because scientists now
hypothesize that comets and asteroids smashing into the lunar surface
eons ago left behind some water. Of course, water on the Moon’s surface
does not last very long. It evaporates in the intense sunlight and quickly
departs this airless world by drifting off into space. Only in the frigid
recesses of permanently shadowed craters do scientists expect to find any
of the water that might have been carried to the Moon and scattered across
the lunar surface by ancient comet or asteroid impacts. In the 1990s, two
spacecrafts, Clementine and Lunar Prospector, collected tantalizing data
suggesting that the shadowed craters at the lunar poles may contain significant
quantities of water ice.
NASA plans to resolve this very important question by using smart
robots as scouts. First into action will be the Lunar Reconnaissance Orbiter
(LRO)—a robot spacecraft mission planned for launch by late 2008. The
LRO mission emphasizes the overall objective of collecting science data
that will facilitate a human return to the Moon. As part of NASA’s strategic
plan for solar-system exploration, a return to the Moon by human beings
is considered a critical step in field-testing the equipment necessary for a
successful human expedition to Mars later in this century.
The LRO will orbit the Moon for at least one year using an 18.6–31.1-
mile- (30–50-km-) altitude polar orbit to map the lunar environment in
greater detail than ever before. The six instruments planned for the Lunar
Reconnaissance Orbiter will do many things: they will map and photograph
the Moon in great detail, paying special attention to the permanently shadowed
polar regions. The LRO’s instruments will also measure the Moon’s
ionizing-radiation environment and conduct a very detailed search for
signs of water-ice deposits. No single spacecraft-borne instrument can
provide definitive evidence of ice on the Moon, but if all the data from the
LRO’s collection of water-hunting instruments point to suspected ice in
the same area, those data would be most compelling.
Within NASA’s current strategic vision for robot-human partnership
in space exploration, the LRO is just the first in a string of smart robots
with missions to the Moon over the next two decades. Once compelling
evidence for the presence of water ice is obtained by the LRO, then the
next logical step is to send a smart scout robot to that location to scratch
and sniff the site and to perform on the spot (in situ) analyses. The rover
robot’s detailed investigations will confirm the existence of any water ice.
The semiautonomous mobile robot may expand investigations of the
area to provide a first-order estimate of the total quantity of the water
available.
Finally, if suitable water resources are located and inventoried, teams
of smart robot prospectors would be sent to the Moon to harvest the particular
site or sites in preparation for the return of human beings to the
lunar surface. Supervised and teleoperated by humans from Earth, a team
of semiautonomous water-harvesting robots would make the construction and operation of a permanent human base practical (from a logistics perspective)
and prepare the way for an eventual human expedition to Mars.