Why don't robotic spacecraft have "general purpose" companions?

Question

TL;DR: Why don't we have minion bots with WD-40 and duct tape, alongside every robotic mission?

I am a software engineer, but I also enjoy and am good with using manual or power tools. Therefore I know you can get a lot done, built or fixed with great tools, OR approximate tools and a lot of time.

For years, I've been watching the space industry as closely as my free time permits. I've been wondering, and failed to answer, why space engineers do not come up with general purpose robotic companions?

For that question is in my backburner, it did not escape my attention when InSight used it's arm to clean solar array, or hammer the mole.

I am also aware of the Canadarm, it's Russian version, and ESA's upcoming ERA. And all, especially ERA looks quite logical and useful.

Now, being an engineer for years, I know if some idea sounds like a no brainer to an outsider of the domain, it is usually due to lack of knowledge or historical context of the outsider.

Then I have here something that appears to be a no brainer to me, and I shall explain it to ask what am I missing, for I was not able to shake the idea that it is a no brainer, as an outsider to space engineering:

Let's imagine an ERA like symmetrical arm robot species, varied for each mission, but same on the main principles;

• they have a central "body", where they have a "toolbox"
• they achieve locomotion much like ERA, on predefined ports, or using suctions
• they have cameras at various spots
• they can be programmed or remote controlled
• hydra ends: they might have smaller arms at each end, for example main arm could hold a piece while small one applies glue, cranks a bolt, etc.

I think it is a perfect time to ask this question, as JWST has just went up with some 300+ single point failures.

So, Imagine JWST, Hubble, or a rover on mars has failed or stuck somehow. And you are personally there, with a toolbox. I feel like you can fix a lot. So why not put a robotic member and remotely control it as a mechanic?

What is preventing every robotic mission from having such a utility companion? Is it just weight? Would it actually not be that useful? Are there technical difficulties that I'm unaware of?

For example, how much of the Hubble servicing missions could have been done remotely with such robots, if we had such robots?

---

Update for Clarification

What I meant for "general purpose" was something like a surgery robot mixed with an assembly line robot, except there would be many bits or attachments in it's toolbox. There could be a brush, one soft and one heavy hammer, a scraper, several drill bits, and some screwdriver bits, one tiny one strong grip tool etc. I imagine that could have helped with the "mole" issue of the InSight, dust over solar panels issue there and many other missions.

For example, Curiosity has a quite multifunctional arm, so why not a standalone version of it? And why it has to carry it's toolbox at the end of the arm? What kind of problems would picking up attachment when necessary from the rover body have caused? It picks and drops sample tubes with an arm, for example.

And as I tried to point out, I know if there's no such thing already, there would be good reasons for it. So, the question is: What am I missing?

• Would it not be practical? If you aware of any papers prospecting utility of robot maintenance, would be great to point out.
• Would the added weight not justify the utility?
• Would such a robot have many points of failure itself?

Answer 1

General purpose repair robots do not yet exist because the costs have not fallen below that of just launching new (upgraded) satellite that you can test and repair on the ground.

The first challenge with a general purpose repair bot is moving orbits. It is expensive to move from the surface of earth (speed 0) to earth orbit 7 kilometers a second. What is not intuitive is that 7 kms has a vector and moving orbit involves changing that vector. As a worst case, if you want to service a polar orbiting vehicle than an equatorial one, you need a burn to kill 7kms North/South and add 7kms East/West which is MORE expensive in fuel than just launching a new satellite East/West from 0.

Famously Hubble (28 degrees) is not reachable from the ISS (51 degrees)with any current manned vehicle, needing 3kms, or about the same as a LEO to moon transfer.

So this means that owners of constellations, the ones with large numbers of identical birds that might be the most likely customer for being serviced do not actually have much use for them unless they put one in each orbital plane, and if doing that it is probably easier to launch complete spare satellites than a robot carrying spare parts for satellites that it then has to fit.

The exception here is geostationary. There are several hundred satellites there, all in the same plane where transfers can be very cheap if you are prepared to wait a couple of weeks. This is why the only proposed commercial service missions (refueling or fitting new engines) have focused on this region.

The second problem is approach. Rocket exhaust is hot, and often corrosive and a simple approach per newtons law means your final 'stop' burn will be towards the target at short range. This is a solvable problem with careful trajectory design and off axis burns but a challenge - and certainly impacts 'hover around' or 'grip onto and pull stuck thing' type concepts. Also means that ideally you want a target that is not tumbling, which limits your servicing to not yet failed vehicles.

Actually doing the service is a challenge as well. The simplest function is refueling an otherwise serviceable craft, which has been done (to the point of being routine for ISS). Replacing parts is where it gets trickier. The first challenge is identifying the problem, since while it might be 'widget A' it might also be any of the myriad of power, cooling, clock or sensors that feed widget A (imagine doing phone support where 10% of the time the person lies to you). Second problem is that space vehicles need to be compact to fit onto the rocket, and survive a great deal of vibration and G force on ascent so they are not generally like a computer case and more like the inside of a very large phone, with cooling plumbing threaded in and all connections screwed, glued or lockwired. And the whole affair covered in thermal film.

So a service that upgraded solar panels might be possible doing an internal part replacement starts to become a complex affair involving many different tools (including glue) and in some cases significant force and it all needs to go back together.

The Hubble repair missions are a useful data point - it was stable during approach, it was designed with many duplicate systems, so that they could be reasonably certain of what was broken and because of the size of the optics many of the electronic parts were in well spaced apart stand alone modules designed with some thought to service (including some on swing out doors) and they were still very challenging missions.

There is the Dextre system, which has been working on the ISS for 10 years but does not appear to have been used over space walking humans for any actual repair type activities, suggesting where the limits of space based robot manipulation currently exist.

So currently robotic repair is generally more expensive than a new mission, with some exceptions (refueling). As launch costs fall it may become more viable to justify things like service doors over fixed panels, and plugs over solder on satellites that make repair more viable, but the fuel cost of moving between orbits will remain a big viability problem.