Skip navigationHumans vs. Robots: Who's best in space?

1. Learning from Columbia

2. A troubled hybrid

3. The argument for people

4. Pointless exercise?

5. Making sense of the debate

 

 

On Mars in 1997, Sojourner took an X-ray measurement to determine the composition of a rock nicknamed "Yogi." The brainy breadbox caught our attention, and for good reason: People won't reach the Red Planet for a couple of decades. Photo: NASA.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

What is the justification for sending astronauts into space to do science?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Moss grown in the low-gravity conditions aboard space shuttle Columbia, in a previous mission. The spiral pattern may tell us lots about how cells grow in space. Photo: NASA.

 

More bad news
The official studies of space science tend to substantiate Robert Park's pessimism, especially concerning the space station. Physicist Herman Cummins, for example, was on a National Research Council committee charged with "evaluating the scientific merit and feasibility of carrying out various planned experiments," on the station. Cummins, who studies liquids, crystals and gases at CCNY, does not launch science experiments into space.

A breadbox on six wheels noses up to a rock many times larger.

The advisory committee, which reported in September, 2002, concluded that robots could handle many projects. "My personal opinion," says Cummins, "is that the amount of human participation necessary for [scientific experiments in] the space program is actually very little. Much of the work could be done by automated systems."

Many experiments, he says, are already under computer control. "The astronauts basically turn a knob once a day, which may not be necessary."

That situation reflects a NASA requirement that experiments involve some human participation, he adds. "The human element makes this very expensive program attractive to the general public, so ... if you say you will operate [an experiment] entirely through radio control, they will say you can't fly that experiment on ISS or the shuttle."

However, we did hear a slightly different story from Tony Ingraffea, a professor of civil and environmental engineering at Cornell University. "Many experiments on the shuttle and space station are almost entirely automated, but on no flights are all of them automated."

Researchers react
How do proponents of doing research on the shuttle and the space station justify the work? Fred Sacks is a professor of plant biology at Ohio State University who saw an experiment on the growth of moss in the micro-gravity burn up on Columbia. What, specifically, would direct the moss to grow upward when there was almost no gravity to indicate "which way was up"?

Sacks was at Cape Canaveral for Columbia's landing, and says, "We knew something was wrong when they were two minutes late. Then NASA said they'd lost contact." The emotional response, he says, was complex. "There were many stages of grieving through the first few days. It was overwhelming. Like everybody else, I was in total shock."

A few days later, a different reality began to sink in, he told us just after his return from Florida. "We're starting to realize how much work we lost." For five years, his crew of six scientists, together with a team at Kennedy Space Center, had built and tested the plant-growth hardware. "It's quite striking to see how complicated, how much work a spaceflight entails," Sacks says. While the equipment apparently worked, he "got no hard data."

Plants grow
For long space voyages, plant growth is key because plants supply food and liberate oxygen from carbon dioxide. Sacks' experiment, however, also had a basic-science goal. "People have not fully explored the impact of gravity on biology and human physiology," he says. Life evolved under Earth's constant gravity, "and the more we look, the more we question our assumptions, the more we will discover about biology."

Several years ago, Sacks says, when neonatal rodents developed on a shuttle, they were "much more disoriented than the control animals. It raises the possibility that gravity itself may be necessary for the development of neural apparatus."

If people are going to take long space missions - to Mars or beyond - we need to limit the damage space travel causes to digestion, muscles, bones and the immune system, Sacks says. "If we're going to have people in space, and I can't imagine humans would not ultimately want to be in space, we need to start studying physiology, how to supply food in space...the basics of how gravity affects life."

Having said all this, however, Sacks, like other scientists, is hesitant to judge the proper balance of the dangers of space-based science and the human cost. "That's a hard question. I think no one experiment can justify [the loss of life], but I think in general, space is a lab that's essential."

Low-gravity moss grows in a whirling pattern, regular moss grows more vertically.

The human element
The benefits of automating research in space extend beyond saving lives, says Ingraffea, the Cornell University engineer. Life-support systems are expensive to design, build, launch, supply and maintain. "You could make a strong argument that the lifetime cost of something that does not have people on it in Earth orbit would be substantially less."

While research into, say, crystal formation could be completely automated, Ingraffea says even self-contained experiments must "be started, monitored, and the data gathered." Further, he adds that years of remodeling would be needed before the space station could handle fully automated experiments.

Perhaps the most significant drawback of full automation, he says, is a loss of flexibility. "What people bring to scientific experiments is the ability to do something different as suggested by some immediate outcome, and that, after all, is the beauty of scientific experiments." Robots, he says, are better at blindly stumbling ahead than changing course to match new conditions.

Indeed, in the face of unexpected experimental results, "Robots are specifically designed to ignore most of those as noise," says Barrett Caldwell, an associate professor of industrial engineering at Purdue University.

Robot woes
Caldwell, who also directs the Indiana Space Grant Consortium, is an expert in human-machine interfaces. By email, he pointed to other limitations of space robots:

"Time delays and limited sensor capabilities drastically impair the remote operator's ability to control a robot.

"Size and weight requirements are still pretty significant.

"[Automation] requires additional manipulation of the experiment to support the (limited) capability of the robot.

"The robot's engineering system adds another level of complexity and failure modes, which may not be fixable by the remote operator."

When it comes to fully autonomous robots - those that need human at the other end of the control system -- Caldwell says, prospects "are much dimmer," since they also must recover from errors and do unplanned tasks.

Counter-Barrety fire
To physicist Robert Park, director of communications for the American Physical Society, these cavils reflect such a misunderstanding of modern robots that they amount to "a red herring." These robots "don't have to make decisions, that's the hardest thing to get across," he insists.

Park points to one of NASA's great successes - and public-relations triumphs - the July Fourth, 1997, stroll of a breadbox-sized robot vehicle on Mars. "Little Sojourner on Mars didn't have any brain to speak of, it took its instructions from somebody on Earth," says Park "The most wonderful thing about Sojourner was that it wasn't [operated by] a couple of thick-headed astronauts. We were all there, we all got to see Mars through the robot's eyes."

Sojourner, he says, was a "telerobot, which is a simple extension of the human; its brain is the brain of the operator. All the robot does is to put our eyes, ears and fingers in places where human beings can't go."

Space science on the space station. Let's hear from the expert reports.

 
  backmore
 
   
 
 

There are 1 2 3 4 5 pages in this feature.
Bibliography | Credits | Feedback | Search

©2003, University of Wisconsin, Board of Regents.