Skip to main content
1 answer
Updated 231 views Translate

Robonautics benchmarking

We have come a long way, we still have further to go. I see a future where we don't risk human lives on space station repair / constuction. When will this be in the forseeable future?
#human-resources #career #robots #nasa

+25 Karma if successful
From: You
To: Friend
Subject: Career question for you


1 answer

Updated Translate

Joseph’s Answer

To some extent, we're already there, at least for construction. All long-range missions to date have been robotic probes, and they regularly fly automatic supply ships to the ISS which can attach themselves under robotic control - so basically we can already do modular space station construction robotically.

Repair is more complicated, granted - but if there's going to be humans on the station anyway, they might as well do the repairs. There's not that much additional risk in conducting repairs to being there in the first place; and there will be an argument for humans in space for a long time, regardless of the capabilities of robotics.

At some point in humanity's future, we will need to leave this planet and settle other worlds. That means humans will have to travel in space eventually, so we need to have humans in space to understand the effects of being in space on the human body, in terms of being in weightlessness and in with greater exposure to cosmic radiation; and to develop systems and technology to better deal with these hazards, so we need humans there to test these things. I don't really see a future where robotics can replace all human spaceflight.

However, regardless of whether it would be completely handed over to robotics, you're asking when robotics will be at a level to conduct all the maintenance tasks a human could do - and that's very difficult to answer. I think we have the capability today to build remote manipulator arms that could replicate all human motion, but with the lag of sending signals to orbit, full ground based human control is not practical for certain tasks - some things need instant feedback with no perceptible lag. That means at least some operations have to be completely autonomous with no human-in-the-loop; which requires an intelligence to make decisions; and that's where the big stumbling block is - we're a long way from human-level artificial general intelligence, and even if we were close, the route there is fraught with AI safety risks.

Perhaps by blending some human control with some automation, we would be able to achieve these sort of tasks, but that can be highly task dependent. The sort of things Boston Dynamics are doing with Spot and their other robots are good examples of the current technology here. For all the tasks needed on a space station, I'd guess those capabilities are still greater than 50 years away.