The Space Reviewin association with SpaceNews

NSRC 2023

telerobotics illustration
A key factor in any future human space exploration effort may be the level of latency—time delay—we’re willing to accept for teleoperated robots on other worlds.

Human spaceflight, and the reason for (almost) being there

Bookmark and Share

In the long debate about human space exploration—how to achieve it, and what it’s for—telerobotics is increasingly important. Can exploration be carried out using “dumb” but sophisticated electromechanical extensions of our dexterity and senses that put humans virtually at a remote site? The virtual presence achieved that way is as “real” as the fidelity with which our dexterity and senses can be electromechanically communicated. For vision and hearing (to the extent the latter is necessary in space), our new technologies can do this relaying, and can easily surpass that of human eyes and ears in doing so. For smell and taste (again, if necessary), sensors are available that can quantitatively distinguish among a large range of substances. Haptic arrays are beginning to allow humans to “feel” things remotely. Precision dexterous manipulators are in regular commercial use. Transcontinental surgery, as we do now remotely, allowing delicate incisions, stitching, and even tying of knots, suggests that telerobotically picking up rocks and turning screwdrivers shouldn’t be that hard. These capabilities are seriously degraded for a human in a spacesuit. So why do we need humans around?

Humans are essential to exploration. However, technology has now advanced to the point that we can wonder how much humans have to physically be there to do it. As emphasized by many commentators, the word “exploration” is a slippery one for space. Historical exploration doesn’t pertain well to modern space exploration, as those historical explorers didn’t have communication, recording, and telerobotic tools that are now widely available. Early generations of explorers had to “be there” physically to “be there” functionally and cognitively.

Humans are essential to exploration. However, technology has now advanced to the point that we can wonder how much humans have to physically be there to do it.

Space scientists have a long history of accomplishment with robotic probes, which swoop through the solar system, down on planets, asteroids, and moons. But the Mars Exploration Rovers (MERs) set a new standard for robotic command of a complex, obstacle-ridden environment. Human space explorers in Pasadena could drive for miles across Martian dunes, around boulders, onto hilltops, into craters, and peer closely at rock formations. Living on “Mars time”, with a 24.7-hour day, the controllers of Spirit and Opportunity found themselves deeply immersed not just in the science but in the discovery process. This was described by Janet Vertisi in her CHI 2008 contribution “Seeing Like a Rover: Embodied Experience on the Mars Exploration Rover Mission” (DOI: 10.1145/1358628.1358709), and Bill Clancey in his 2009 essay “Becoming a Rover”. The MER scientists were convinced that their work embraced the spirit of exploration, using machines as surrogates. Every day they were looking, searching, testing, questioning… and exploring.

But as capable as the MERs were, they were hardly optimal for discovery. While vastly more expensive, humans on site could provide a wealth of discovery more quickly. MER Principal Investigator Steve Squyres, in a spirited defense of human missions to Mars, memorably noted that the MERs would take a day to do what a skilled field geologist could do in minutes. While the robotic systems on Spirit and Opportunity were primitive by modern standards, Squyres’ point is quantitatively justifiable especially with regard to just moving around. The reason is simple: communication latency. The two-way time delay (latency, lag, or ping time) between the Earth and Mars is at least eight minutes, and as many as forty, and the sense/question/respond/confirm cycle of discovery with Earthbound humans in the loop is fundamentally limited to this timescale. While human operators are virtually on Mars, their sense of “presence” is profoundly compromised. The time delay makes for a poor cognitive connection, and a limited sense of “being there”. When you want to look at a rock, you send the command to do it, and go out to have lunch before coming back to find out what the rock looks like. That’s when you’d really like a better sense of presence! A Turing-like test of telerobotics for space exploration might be whether “presence”-wise a controller would feel like he or she was actually there.

While sensory and dexterous capabilities of robots are going to improve enormously in coming decades (and already have since the MER Critical Design Review a decade ago), communication delays will not. We can’t expect any increases in the speed of light. To the extent that we want to do exploration with a true cognitive coupling of human conceptualization and perception to a task, as opposed to autonomous exploration where we have to “trust” a robot to make the right judgments, distance makes life hard.

We come to the remarkable conclusion that, in a technologically sophisticated era with advanced robotic capabilities, the ultimate practical purpose of human spaceflight, for whatever task, may be to minimize latency. Such minimization establishes a high quality cognitive connection between a human brain and the site to be explored. In the words of the MIT Space, Policy, and Society Research Group, human space exploration is broadly aimed at “an expansion of human experience”, where that experience is best expanded by bringing the human brain close enough to the site. While expanding human experience is a clear mandate for science, it also bears on resource development, and the concomitant need for mining and refining equipment to be constructed, controlled, and maintained. It also relates to what can be termed “conquest”, in which cognitive connection establishes presumptive control over a site, much as we have in Afghanistan with remotely operated drones. With requisite telerobotic sophistication, lowering latency puts the “presence” in virtual presence.

We come to the remarkable conclusion that, in a technologically sophisticated era with advanced robotic capabilities, the ultimate practical purpose of human spaceflight, for whatever task, may be to minimize latency.

The obvious question then is, how close do human operators need to be? To a high-performance online gamer, a two-way latency of 100 milliseconds (15,000 kilometers each way, by the speed of light) can be problematical. But for transcontinental telerobotic surgery, up to 500 milliseconds is considered endurable by surgeons, most of which would be network-processing time. Similar latencies are common in the rapidly expanding field of telerobotic mining, in which bulldozers and drills are driven by office workers thousands of miles away. Longer latencies are cognitively more difficult than when they are shorter, as the brain works in a different way, but latencies of 1 to 2 seconds are not uncommon for military drone piloting. A one-way latency of about a quarter of a second multiplied by the speed of light is about 75,000 kilometers and roughly defines, as Lester and Thronson assert in their recent Space Policy paper (DOI: 10.1016 /j.spacepol.2011. 02.002) what we can call the cognitive scale of our universe.

The latency needed to achieve a high quality cognitive connection with a robot establishes a policy foundation for human space exploration. That limiting latency bears on the “flexible path” strategy espoused in the Augustine Report. A key tenet of that strategy is that exploration can be done more quickly and economically if we can avoid, at least in the near term, sending humans into deep gravity wells. Those gravity wells demand development of complex human-rated ascent/descent vehicles. Additionally, traveling to, landing on, and operating across a planetary surface requires two complete, and almost unrelated complex life support systems. But we don’t have to dive into those gravity wells to achieve high quality cognitive connection with things on the surface. We just need to get close.

Lester and Thronson have suggested here (see “Human operations beyond LEO by the end of the decade: An affordable near-term stepping stone”, The Space Review, January 10, 2011) and in their new Space Policy paper (see above) that lunar surface operations can be telerobotically controlled with a high degree of cognitive connection from the first and second Earth-Moon Lagrange points (about 40,000 kilometers from the lunar surface), where stationkeeping cost is low, visibility is continuous, and there are many additional reasons to have humans at a facility there. Similar strategies have been proposed (e.g. “Human Exploration using Real-time Robotic Operations” – HERRO) for the Martian surface—a challenging, and intensely interesting environment—using telerobotic control, perhaps from Phobos, Deimos, or low Martian orbit.

If species expansion is the ultimate goal, we can’t do that with robots.

In principle, such low-latency on-orbit control architecture can be used to explore, with very real sense of human “presence”, surface locations far less hospitable to human life. Scouting mountains on Venus? Undersea diving in the oceans of Europa? Sailing across the lakes of Titan? Leaning over the caldera of an Ionian volcano? While landing on the Moon may help us develop architectures that will help us land on Mars, on-orbit control of lunar surface telerobots is an exploration strategy extensible across the entire solar system. It dramatically increases the list of potential sites for human exploration. Although not specifically relevant to human space flight, perhaps the first application of low latency telerobotics in space will be the rescue, refueling, or servicing of expensive satellites in geosynchronous orbit from dedicated ground stations, through a round-trip latency of about 250 millseconds.

Telerobots should be an economical approach to many surface tasks. They can be conveniently placed and controlled at multiple surface sites. At least for the Moon, for science, ISRU testing, and even for outpost development and emplacement, control responsibility can be shared with the Earth, 2,600 millseconds away round-trip. Tasks demanding six times lower latency than from the Earth can be controlled from Earth-Moon L1 or L2. Lower latency is possible from low lunar orbit, but at additional operational costs.

This is certainly not to say that humans shouldn’t venture into gravity wells, for example down onto the surface of the Moon or Mars. While the sophistication of our telerobots is likely to improve dramatically, it is hard to say when they would substitute entirely for human presence. Furthermore, there are historical images of exploration that are well served or at least honored, in the public eye, by having astronauts planting their feet on other worlds. Finally, if species expansion is the ultimate goal, we can’t do that with robots.

One can anticipate some discomfort in human space flight advocates about on-orbit telerobotic exploration. “Looking without touching” has been a common complaint about the “flexible path” strategy. We suggest that with sufficiently sophisticated telerobots, and a latency that permits a real sense of virtual presence, the experience of on-site exploration can be quite real. Arguably more real than to an astronaut entombed in a bulky and constraining space suit.

“But who fixes the robots?” is another common lament. The answer is simple: probably other telerobots. Fixing a robot that offers extension of human senses and dexterity allows us to presume that there is another one that can bring human senses and dexterity to bear in fixing the first one. To this question might be added a parallel one. “Who fixes the people?” For a given support cost, it’s likely that people will need fixing more than robots. In fact, disabled humans would be replaced, as we would do for a seriously malfunctioning telerobot. Except for the robot we don’t need to bring back the sick one.

In the same way that our science and technologies have developed, our presumptions and preconceptions about exploration should mature as well.

“I know robots, and they aren’t anywhere near that sophisticated” is another criticism we can anticipate. There are several ways to address this. As pointed out above, sophisticated telerobotic surgery is becoming common, and telerobotic mining is actually a major cost-saving industry strategy. These technologies would need upgrades to obviate the need for proximal human supervision, which is easily available on Earth. Advances in such low-latency telerobotics are not only happening very rapidly, but they are being driven and funded commercially for terrestrial applications. On a decadal timescale of human space flight, major advances can be confidently expected that will better enable virtual presence.

Finally, one might ask, “If telerobots are so good, why aren’t we using them everywhere now on Earth?” That’s easy. Because humans on the Earth are dirt cheap. Humans in space, and especially on the surfaces of other worlds, are not.

Philosophically, this essay essentially asks what constitutes human space exploration. As noted by Lester and Robinson, that phrase is uncomfortably flexible and ill-defined (DOI: 10.1016/j.spacepol.2009.07.001). If human space exploration is primarily about getting toes dirty, then on-orbit telerobotics won’t suffice. If it is task-driven, about investigation, discovery, maintenance, and construction, needing a high degree of human cognitive coupling to do those tasks, then telerobotics can be an economical approach, and can be achieved in a quality way by having human operators almost there. It is a compelling vision for space exploration that benefits from both human-robot partnership and human space flight. It touches on what can be called skeptical questions, such as whether telepresence is really equivalent to proximal experience, a matter that is known as technical telepistemology. While telerobotic technology is in its infancy, it is inescapable that such sophisticated capability will eventually be well realized. The challenge is how well the public can accept human space exploration of a venue by humans who aren’t quite all the way there. Who’s going to buy it? This, too, may provoke a sense of unease in human space flight advocates, in that the very premise of “destination” is called into question. But in the same way that our science and technologies have developed, our presumptions and preconceptions about exploration should mature as well. This maturation is a major challenge that the enterprise presents to us.