The people to ask, really, are those who planned and operated the Lunakhods back in the '70s. These were the only rovers ever landed on an extra-terrestrial body that were operated in a low-latency, run-it-with-a-joystick mode. From the simple measure of distance traveled divided by time, we can obviously see that the Lunakhods had a much greater capability of traveling longer distances over shorter timeframes -- one of them was able to cram 30-plus km of travel in just a few months. Compare this to the several years it took for Oppie to travel the same distance.Unfortunately, the people who designed the operations processes for the Lunakhods, and the people who actually operated them, are mostly long-dead (or at least long-retired); I'd have doubts about being able to pull out their lessons learned, or why they decided on such a low-latency paradigm vs. the plan-a-day's-worth-of-operations-and-let-'er-go paradigm we've used for our American Mars rovers. It would be really interesting to see or hear any oral histories captured from those people.Now, total travel distance over time is not necessarily the best metric for how to design tele-operations. But it is one area where low-latency operations make a difference. Again, though, it would be very interesting to see a rigorous study done to analyze both paradigms and see what might be gained by deploying tele-operators from Mars orbit and operating them in low-latency modes.
And again, I note that the value of low-latency telerobotics is usually asserted, but not actually studied or demonstrated. People say "of course you want to eliminate the time delay," without looking carefully at each operation (driving, moving an arm, operating a science payload, surveying terrain) and asking what is the difference between low-latency and high-latency for that operation.
Quote from: the_other_Doug on 05/23/2016 02:09 pmYep. And JPL seemed to say "Thanks for your information, but it was Not Invented Here, so we won't be using any of it."And it wasn't so much the Lunokhod/Marsokhod designers, it was the designer, singular. A guy named Alexander Kemurdzhian. Yes, he had a design team in the USSR back in the '60s, but I'm pretty certain he was the only person invited to "compare approaches on the design of planetary rovers" with designers at JPLIIRC, he brought a small test rover he had been playing with and demonstrated it for the JPL people, but again, this one didn't have any design heritage that was built into later JPL-designed rovers. And from what I've seen of JPL people describing this encounter, they seemed more interested in the design aspect of the Soviet hardware -- not so much the operational paradigm, which is what I was suggesting we check into and see if there's any documentation about.I think Kemurdzhian's biggest claim to fame, outside of Lunokhod, was the fact that he was able to put together, on extremely short notice, remote-operated rovers used to survey the damage on the roof of the reactor building at Chernobyl after the explosion and fire there.Unfortunately, Kemurdzhian has been gone for more than a decade, and I'm pretty certain most of the people who operated the Lunokhods are no longer with us, either -- it's been coming up on half a century since they were active, after all. So, that one encounter JPL had with Kemurdzhian was pretty much it, in terms of checking with those who operated the only real-time rovers ever operated off-Earth.I'd love to see oral histories captured at that time that give details on rover operations discussions JPL may have had with Kemurdzhian, but I've never seen such come to light -- and, just to point out the obvious, the people at JPL who talked with Kemurdzhian back in the '90s are likely now, too, getting long in the tooth, are gone, or are retired. Think about the scene in "The Martian" where they had to gather up surviving members of the Pathfinder team, and recall thinking to yourself "I bet there wouldn't be that many survivors of that team by the 2030s..." Thats quite a limited view of the history. No, it wasnt A. Kemurdzhian alone. No, it wasnt a demonstration to 'JPL people' - Marsokhod based field campaigns were mainly run by NASA Ames and McDonnell Douglas, with instigation by Planetary Society, and continued well into 1999.And, the field tests of Marsokhod and multiple other rover designs from different NASA field centers ended up contributing mostly exactly to operational side of missions like Pathfinder and MERs. There were simulated field tests with long time delays as well. Reports are available on NTRS and referenced in other publications - for instance, Marsokhod fields tests are referenced in landing site workshops for Pathfinder.The hardware itself was subject to multiple iterations of mission proposals like Mars Together, which all died for multiple reasons.
Yep. And JPL seemed to say "Thanks for your information, but it was Not Invented Here, so we won't be using any of it."And it wasn't so much the Lunokhod/Marsokhod designers, it was the designer, singular. A guy named Alexander Kemurdzhian. Yes, he had a design team in the USSR back in the '60s, but I'm pretty certain he was the only person invited to "compare approaches on the design of planetary rovers" with designers at JPLIIRC, he brought a small test rover he had been playing with and demonstrated it for the JPL people, but again, this one didn't have any design heritage that was built into later JPL-designed rovers. And from what I've seen of JPL people describing this encounter, they seemed more interested in the design aspect of the Soviet hardware -- not so much the operational paradigm, which is what I was suggesting we check into and see if there's any documentation about.I think Kemurdzhian's biggest claim to fame, outside of Lunokhod, was the fact that he was able to put together, on extremely short notice, remote-operated rovers used to survey the damage on the roof of the reactor building at Chernobyl after the explosion and fire there.Unfortunately, Kemurdzhian has been gone for more than a decade, and I'm pretty certain most of the people who operated the Lunokhods are no longer with us, either -- it's been coming up on half a century since they were active, after all. So, that one encounter JPL had with Kemurdzhian was pretty much it, in terms of checking with those who operated the only real-time rovers ever operated off-Earth.I'd love to see oral histories captured at that time that give details on rover operations discussions JPL may have had with Kemurdzhian, but I've never seen such come to light -- and, just to point out the obvious, the people at JPL who talked with Kemurdzhian back in the '90s are likely now, too, getting long in the tooth, are gone, or are retired. Think about the scene in "The Martian" where they had to gather up surviving members of the Pathfinder team, and recall thinking to yourself "I bet there wouldn't be that many survivors of that team by the 2030s..."
Thanks for the clarification. I was mostly going off what I've read about Kemurdzhian, and fairly extensive a biography I saw on TV of the man, which focused both on the Lunokhod achievements and the Chernobyl rovers. It had a section in which, during one of the last interviews the man gave, he expressed frustration that no one here in America seemed to take much interest in his work. That interview mentioned the NIH attitude that was frustrating him a lot, and the lack of any follow-up after his demonstrations. (It featured some nice footage of him demonstrating the hardware he brought for the NASA people to look at, though.)It's always good to get additional points of view on any subject. Thanks for this one!
Unfortunately, the people who designed the operations processes for the Lunakhods, and the people who actually operated them, are mostly long-dead (or at least long-retired); I'd have doubts about being able to pull out their lessons learned, or why they decided on such a low-latency paradigm vs. the plan-a-day's-worth-of-operations-and-let-'er-go paradigm we've used for our American Mars rovers. It would be really interesting to see or hear any oral histories captured from those people.
Quote from: Blackstar on 05/23/2016 06:48 pmAnd again, I note that the value of low-latency telerobotics is usually asserted, but not actually studied or demonstrated. People say "of course you want to eliminate the time delay," without looking carefully at each operation (driving, moving an arm, operating a science payload, surveying terrain) and asking what is the difference between low-latency and high-latency for that operation.Tele-robotics is not some unknown or very novel field of study. It's performing commercial, military, medical and research applications on earth at increasingly wider scale. In air, under water, underground, on field and on battlefield. The achievable capabilities are kind of well understood.
Quote from: savuporo on 05/23/2016 07:06 pmQuote from: Blackstar on 05/23/2016 06:48 pmAnd again, I note that the value of low-latency telerobotics is usually asserted, but not actually studied or demonstrated. People say "of course you want to eliminate the time delay," without looking carefully at each operation (driving, moving an arm, operating a science payload, surveying terrain) and asking what is the difference between low-latency and high-latency for that operation.Tele-robotics is not some unknown or very novel field of study. It's performing commercial, military, medical and research applications on earth at increasingly wider scale. In air, under water, underground, on field and on battlefield. The achievable capabilities are kind of well understood.IMHO, latency is only half the issue. The other is bandwidth. The current rovers spend a lot of time sitting, charging their batteries and transmitting data to Earth (direct and/or via relay) between action plans. Sending that data to orbiting operators should be at least an order of magnitude faster, assuming the orbital asset has significantly more buffer space than existing spacecraft like MRO. The local geologist gets to see the data right away and inform the programming team regarding next targets, etc. I would guess 2-3X more science data gathered per sol with little change to existing rover operations.
For field geology, we have NO experience with low latency telerobotics. Lessons from high latency telerobotics (MER, MSL) donʼt necessarily transfer well to low latency telerobotics. Analog studies on the Earth will be essential to this understanding.
QuoteFor field geology, we have NO experience with low latency telerobotics. (from slide 11)
For field geology, we have NO experience with low latency telerobotics.
If a teleoperated rover were to transmit real-time video to orbit, whilst driving vastly greater distances each day than current rovers, how would it be powered? The MERs got enough solar power each day to stay alive and drive a couple of hundred metres. Hard to see anybody fitting a vastly greater area of PVA on a rover. And RTGs aren't exactly low cost.
The current rovers spend a lot of time sitting, charging their batteries and transmitting data to Earth (direct and/or via relay) between action plans.
And any rover you send to Mars is still going to spend a lot of time sitting and charging batteries, or sitting in darkness. So again the question is what is the specific value of having the human right nearby? If the rover is in darkness 50% of its day, then that's 50% of the time you're not using it. If it has to charge its batteries for 25% of the daylight (WAG) then that's now 75% of the day that the nearby human will not be commanding it. Figuring out the value of having a human control it with a low time lag, vs. a team of humans controlling it back on Earth, is important to answering the question if this is a worthwhile thing to do.
Figuring out the value of having a human control it with a low time lag, vs. a team of humans controlling it back on Earth, is important to answering the question if this is a worthwhile thing to do.
Its not that hard to figure out, should anyone actually be interested.
Quote from: savuporo on 05/24/2016 01:34 amIts not that hard to figure out, should anyone actually be interested. So you claim. If it's so easy, how come nobody has done it?
And we're talking about rovers that will cost $100s of millions of dollars and are not analogous to robots used on Earth. For instance, no need to disarm bombs on Mars.
Thats quite a limited view of the history. No, it wasnt A. Kemurdzhian alone. No, it wasnt a demonstration to 'JPL people' - Marsokhod based field campaigns were mainly run by NASA Ames and McDonnell Douglas, with instigation by Planetary Society, and continued well into 1999.And, the field tests of Marsokhod and multiple other rover designs from different NASA field centers ended up contributing mostly exactly to operational side of missions like Pathfinder and MERs. There were simulated field tests with long time delays as well. Reports are available on NTRS and referenced in other publications - for instance, Marsokhod fields tests are referenced in landing site workshops for Pathfinder.The hardware itself was subject to multiple iterations of mission proposals like Mars Together, which all died for multiple reasons.
And, NASA itself and ESA have been running various teleoperation programs and field tests for about two decades now, including simulated time lag vs direct operated systems even going as far back as the Marsokhod mentioned above, if Mars geologists do not know what the comparative capabilities are then maybe thats easy to fix with a lecture.
But it's the value of doing this as part of a larger system, including the costs...You're saying that based upon terrestrial examples this is easy and obvious, and I'm saying that from a human spaceflight standpoint and a planetary robotics standpoint none of that has been proven, nor is it obvious.