Author Topic: A limiting factor in robotic planetary exploration  (Read 15650 times)

Offline sivodave

  • Full Member
  • *
  • Posts: 164
  • Liked: 1
  • Likes Given: 0
Hi all.

I was thinking about the two rovers currently on Mars, Spirit and Opportunity. In five years of exploration of the martian soil, these two small rovers have driven no more then a couple of miles. Why? In my opinion the answer is that the communication delay with Earth (20 minutes for sending a message, other 20 for receiving the answer) doesn’t allow for a real time piloting of these machines. This idea of course applies to any kind of robotic exploration you want to do on the surface of any celestial body in the solar system.

What do you think?

Davide

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37440
  • Cape Canaveral Spaceport
  • Liked: 21450
  • Likes Given: 428
Re: A limiting factor in robotic planetary exploration
« Reply #1 on: 07/19/2010 11:18 pm »
no, with AI they can drive themselves.

Offline ugordan

  • Senior Member
  • *****
  • Posts: 8520
    • My mainly Cassini image gallery
  • Liked: 3543
  • Likes Given: 759
Re: A limiting factor in robotic planetary exploration
« Reply #2 on: 07/19/2010 11:30 pm »
Power is the limiting factor, not communications delay.

Offline KelvinZero

  • Senior Member
  • *****
  • Posts: 4286
  • Liked: 887
  • Likes Given: 201
Re: A limiting factor in robotic planetary exploration
« Reply #3 on: 07/20/2010 09:17 am »
So how fast would NASA drive teleoperated rovers on mars with that lag, if power was not an issue? No particular limitation? Just limit to hops that the camera has given us a fairly good look at?

Offline racshot65

  • Senior Member
  • *****
  • Posts: 2577
  • Aaron Kalair
  • Coventry, England
  • Liked: 0
  • Likes Given: 0
Re: A limiting factor in robotic planetary exploration
« Reply #4 on: 07/20/2010 10:28 am »
Talking Space Interviewed a Mars Rover Driver

http://spacetweeps.podbean.com/2010/06/03/episode-219-scott-had-the-spirit-to-accept-opportunity/

And theres a blog post here about speeding up opportunities drive to the Endeavour Crater

http://roadtoendeavour.wordpress.com/2010/07/17/pedal-to-the-metal/

Offline ugordan

  • Senior Member
  • *****
  • Posts: 8520
    • My mainly Cassini image gallery
  • Liked: 3543
  • Likes Given: 759
Re: A limiting factor in robotic planetary exploration
« Reply #5 on: 07/20/2010 12:45 pm »
No particular limitation? Just limit to hops that the camera has given us a fairly good look at?

Best case scenario, meaning freely traversable terrain, yes, you could do large hops at once and depending on your downlink to Earth frequency you could do that a couple times a day.

Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 39270
  • Minnesota
  • Liked: 25240
  • Likes Given: 12115
Re: A limiting factor in robotic planetary exploration
« Reply #6 on: 07/20/2010 03:42 pm »
The DARPA Grand Challenge 2005 has shown that, if given enough power, autonomous vehicles can go pretty far rather quickly on "off-road" terrain (i.e. 150 miles in 7 hours), although there are obviously differences in operating on Mars and the constraints of that challenge. For instance, I believe it was known before-hand that the path of the challenge was traversable, and more risks can be taken if you're in a race with a $30,000 vehicle versus on a unique multi-year mission costing billions.

http://en.wikipedia.org/wiki/DARPA_Grand_Challenge#2005_Grand_Challenge
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Offline GClark

  • Full Member
  • ***
  • Posts: 377
  • Liked: 55
  • Likes Given: 5
Re: A limiting factor in robotic planetary exploration
« Reply #7 on: 07/20/2010 04:36 pm »
Not to mention that this is Science.  They frequently stop to do just that.  The drives are short mostly because of power, but they also drive to a spot they want to spend some time at.

The Lunokhods may have driven further, but the MERs are stopping more often to do intensive studies.

Offline sivodave

  • Full Member
  • *
  • Posts: 164
  • Liked: 1
  • Likes Given: 0
Re: A limiting factor in robotic planetary exploration
« Reply #8 on: 07/21/2010 12:34 am »
Well but i think that the communication delay poses some kind of limitation for a robotic exploration mission. maybe is not a big issue but relevant somehow.

Another thing: if we want to explore Mars (or anyother body of the solar system) we can think to explore it inch by inch otherwise it'll take ages. I think that once an area has been explored, the rover should be moved to another area...of course assuming we have enough power.

it's also true that on Mars there are different type of terrain and till now every rover sent to Mars has been sent in a quite easy area to drive. If we want to explore the planet to its full we have also go in more rough areas and then it could be more difficult drive a rover with a time delay. 

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #9 on: 07/21/2010 03:52 am »
Well but i think that the communication delay poses some kind of limitation for a robotic exploration mission. maybe is not a big issue but relevant somehow.
Of course it's a factor. It just isn't a fundamental limiting factor for the kind of missions we can actually build.

I think ugordon actually overstated the case a little bit (even though the basic point is right on.) The MERs are power limited, but the power would be used more efficiently with real time control. A human would have noticed purgatory dune was a problem much sooner, and the following sols (and watt hours) would have been put to driving instead of digging out. Same goes for many other wasted sols.

On the other hand, MERs communication round trip time is much longer than light time. The rover has to have a scheduled pass with an orbiter (Odyssey is the main relay, has an orbital period of ~2hrs, not every orbit is usable), the orbiter has to relay to earth while fitting into the busy DSN schedule, the team has to analyze the data and so on. If you were willing to invest enough in communications infrastructure, you could operate a rover on Pluto on the same kind of schedule the MERs are on. Wouldn't that be cool ;)
Quote
Another thing: if we want to explore Mars (or anyother body of the solar system) we can think to explore it inch by inch otherwise it'll take ages.
It's an entire planet. Exploring it will take ages.
Quote
I think that once an area has been explored, the rover should be moved to another area...of course assuming we have enough power.
Which we don't, and won't for the foreseeable future.

Quote
it's also true that on Mars there are different type of terrain and till now every rover sent to Mars has been sent in a quite easy area to drive. If we want to explore the planet to its full we have also go in more rough areas and then it could be more difficult drive a rover with a time delay. 
So we will need other techniques to explore those areas. *shrug* Wheeled rovers aren't the only option, and autonomous capability is advancing very rapidly. And again, it's an entire freakin' planet, so even if some places are off limits, we aren't about to run out of interesting things to explore.

Offline docmordrid

  • Senior Member
  • *****
  • Posts: 6334
  • Michigan
  • Liked: 4207
  • Likes Given: 2
Re: A limiting factor in robotic planetary exploration
« Reply #10 on: 07/21/2010 05:23 am »
A tele-operated Robonaut 2 or 3 on the surface of an otherwise human-inhospitable world, but with humans in orbit for short command relays, would give most of the advantages of a human landing with far reduced risks - remembering that humans won't fell much tactically anyhow due to thick gloves and stereo HD cams give darned good remote 3D vision. 

Heck; with today's work in prosthetics, a lot of it at DARPA, sensors can give synthetic remote senses of touch, pressure and texture beyond what an astronaut could feel with gloves on, plus force feedback.  The DARPA arm, which has 22 degrees of freedom and an implanted brain-machine interface, goes into FDA human trials very soon.
« Last Edit: 07/21/2010 05:32 am by docmordrid »
DM

Offline simonbp

  • Science Guy
  • Senior Member
  • *****
  • Posts: 7138
  • Liked: 314
  • Likes Given: 183
Re: A limiting factor in robotic planetary exploration
« Reply #11 on: 07/21/2010 06:55 am »
Power is the limiting factor, not communications delay.

Nope. It's not power. It's the darn scientists.

Even at their best, the rovers only drove at ~20% of their theoretical speed, because the faster they go, the more interesting things they see. And, since most of the instruments took time to measure whatever they saw, any "interesting thing" adds considerable non-moving time.

And thus the difference between space engineers and scientists: the engineers want to go everwhere fast (see "flexible path"), while the scientists want to go to one or two places and actually understand them.

Offline Archibald

  • Senior Member
  • *****
  • Posts: 2611
  • Liked: 500
  • Likes Given: 1096
Han shot first and Gwynne Shotwell !

Offline sivodave

  • Full Member
  • *
  • Posts: 164
  • Liked: 1
  • Likes Given: 0
Re: A limiting factor in robotic planetary exploration
« Reply #13 on: 07/21/2010 08:39 am »
Quote
I like this
http://www.nasa.gov/pdf/372848main_HERRO%20abstract%20GRS.pdf

:)

ah I thought I was the first one to come up with this concept http://forum.nasaspaceflight.com/index.php?topic=22053.0  :-[

I've to think through something different then  ;D

anyway it seems an interesting concept, and maybe more easily feasible then going head-on straight towards a Mars landing.

Offline Garrett

  • Full Member
  • ****
  • Posts: 1134
  • France
  • Liked: 128
  • Likes Given: 113
Re: A limiting factor in robotic planetary exploration
« Reply #14 on: 07/21/2010 10:37 am »
Power is the limiting factor, not communications delay.
Nope. It's not power. It's the darn scientists.
Well, one would presume that there are many factors, but power and the "slow" scientists probably make up the bulk of the "issue". It is of course only an issue to the non-scientist who is more concerned with seeing a lot of lovely Mars terrain imagery. A planetary scientist would probably be just as happy to send a static probe, with the mass of the roving equipment (motor + wheels) replaced with more instruments. The holy grail, so to speak, would be to return Martian soil to Earth and analyse it with higher performance instruments. I would presume that is a much higher priority than making rovers travel faster.

« Last Edit: 07/21/2010 10:38 am by Garrett »
- "Nothing shocks me. I'm a scientist." - Indiana Jones

Offline ugordan

  • Senior Member
  • *****
  • Posts: 8520
    • My mainly Cassini image gallery
  • Liked: 3543
  • Likes Given: 759
Re: A limiting factor in robotic planetary exploration
« Reply #15 on: 07/21/2010 12:20 pm »
Power is the limiting factor, not communications delay.

Nope. It's not power. It's the darn scientists.

It's power. Case in point: Opportunity trek toward Endeavour crater. Of course I'm assuming there aren't deliberate stops for science as I believe the OP was implying that.

The rovers weren't sent there to max out the odometer, they were sent there to do science.

Offline Hungry4info3

  • Full Member
  • ****
  • Posts: 525
  • Liked: 163
  • Likes Given: 81
Re: A limiting factor in robotic planetary exploration
« Reply #16 on: 07/21/2010 12:58 pm »
It's definitely power.

Up until just recently, it's been the same reason they haven't been able to drive every Martian day.  They would drive 50 metres, then recharge the next day.

Offline simonbp

  • Science Guy
  • Senior Member
  • *****
  • Posts: 7138
  • Liked: 314
  • Likes Given: 183
Re: A limiting factor in robotic planetary exploration
« Reply #17 on: 07/21/2010 06:53 pm »
At this point in the mission, yes, it's power due to encrusted panels and over-cycled batteries. But for the first Martian year, when most of the traversing was done by both rovers, it was the scientists spotting things of interest.

And I'm not just making this up, it was told be me personally by Phil Christensen, the PI of Mini-TES, and one the darn scientists himself...

Offline ugordan

  • Senior Member
  • *****
  • Posts: 8520
    • My mainly Cassini image gallery
  • Liked: 3543
  • Likes Given: 759
Re: A limiting factor in robotic planetary exploration
« Reply #18 on: 07/21/2010 07:26 pm »
And I'm not just making this up, it was told be me personally by Phil Christensen, the PI of Mini-TES, and one the darn scientists himself...

Look, this isn't about deliberate stops where scientists wanted to check something out every minute, it's about the practical limitation to the maximum drive distance is. That neglects any deliberate stops. That limitation was and still is (dust or no dust on the solar panels) the available power generated by the solar panels. For MSL it will be the power generated by the RTG over a day and stored in the batteries.

Offline HIPAR

  • Member
  • Full Member
  • ****
  • Posts: 585
  • NE Pa (USA)
  • Liked: 1
  • Likes Given: 0
Re: A limiting factor in robotic planetary exploration
« Reply #19 on: 07/21/2010 08:04 pm »
It's more about size relating to mobility.

The rovers go slowly because they don't have enough ground clearance to avoid becoming stuck in the sand.  They also need bigger wheels that don't sink up to the hubs.  Of course, they were sized to the delivery constraints.

Spirit is currently mired. Opportunity spent a prolonged session stuck in the sand.
So drivers need to incrementally choose short routes and that's what consumes time.  If they were all terrain vehicles they could drive with more impunity.

---  CHAS

Offline khallow

  • Extreme Veteran
  • Full Member
  • ****
  • Posts: 1954
  • Liked: 8
  • Likes Given: 4
Re: A limiting factor in robotic planetary exploration
« Reply #20 on: 07/22/2010 02:09 am »
As an aside, would these rovers have been designed the same way, if the communication delay were a mere 2.5 seconds round trip? You can speak of power limits or laggard scientists, but the communication delay effected the entire mission's design. There was no need for a high powered rover when the power couldn't be used for driving faster. There's no reason for scientists to hurry when the rover isn't going anywhere fast.
Karl Hallowell

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37440
  • Cape Canaveral Spaceport
  • Liked: 21450
  • Likes Given: 428
Re: A limiting factor in robotic planetary exploration
« Reply #21 on: 07/22/2010 11:37 am »
The rovers drive themselves.  they use AI to steer around objects. 

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #22 on: 07/22/2010 10:05 pm »
The rovers drive themselves.  they use AI to steer around objects. 
Errr... sometimes. The MERs are capable of various levels of autonomous driving, but generally more autonomous it is, the slower and less efficient it gets. When the drivers have a enough knowledge of the terrain to say go from A to B on this exact course they do. Even "autonomous" drives involve a lot of planning.

For anyone wanting to get an idea how the rovers are operated I suggest driver Scott Maxwells blog http://marsandme.blogspot.com/

You can speak of power limits or laggard scientists, but the communication delay effected the entire mission's design.
Certainly true. A real time MER would be a different beast.
Quote
There was no need for a high powered rover when the power couldn't be used for driving faster.
I don't see how you could really make a much higher power MER. Maybe optimistically 2x with solar panels ? That would still be pretty power limited. RTGs aren't much better. Other isotope (or reactor) driven systems could do it.
Quote
There's no reason for scientists to hurry when the rover isn't going anywhere fast.
Not really true. The MER planning cycles, especially in the prime mission ran on very tight schedules.

The planning process is much larger than the communication delay. If real time control were available, some of this would be unnecessary, but there would still be a lot of time put into making sure the vehicle stayed safe and picking the science targets.

Long drives from A to B might be done more like Lunokhod, but OTOH, Lunokhod 2 was probably lost due to driver error.

Offline khallow

  • Extreme Veteran
  • Full Member
  • ****
  • Posts: 1954
  • Liked: 8
  • Likes Given: 4
Re: A limiting factor in robotic planetary exploration
« Reply #23 on: 07/24/2010 06:33 am »

I don't see how you could really make a much higher power MER. Maybe optimistically 2x with solar panels ?

That's a huge boost in power right there. But glancing at the transportation system, I'm puzzled. I see that it supposedly takes 100 watts to run, but only goes 5cm/s at maximum speed. Given the mass of the rover and at maximum power, it should be able to rise vertically at up to three times that speed with perfect efficiency (level surface travel should be much faster than that). That indicates to me remarkable power inefficiencies in the propulsion system for the vehicle. In other words, we don't even need to increase the power in order to go faster.

Quote
Quote
There's no reason for scientists to hurry when the rover isn't going anywhere fast.
Not really true. The MER planning cycles, especially in the prime mission ran on very tight schedules.

The planning process is much larger than the communication delay. If real time control were available, some of this would be unnecessary, but there would still be a lot of time put into making sure the vehicle stayed safe and picking the science targets.

I speak of having a "planning process" that can respond in seconds to an opportunity (such as spotting a dust devil or other transient weather  phenomena) or do something while the bureaucratic wheels are spinning. Things that a "very tight schedule" cannot compensate for.

And merely adhering to an aggressive schedule doesn't mean you're hurrying. After all, they're "making sure the vehicle is safe" and "picking science targets", meaning they're spending time doing other things than fulfilling mission objectives. A short communication delay is an obvious benefit to reducing the risk to an investment of a couple hundred million dollars. You cut reaction times to threats to the mission. That means you can work somewhat faster with the same level of safety and reliability.

And "picking science targets" depends on a feedback loop that apparently is hours long. There's opportunity costs to make a decision under those circumstances that wouldn't hold for a fast communication delay. For example, suppose you start to grind a sample from rock A while some interesting phenomena is observed elsewhere (say you decide nearby rock B is a better choice for the grinding sample). Deciding to move under the current scheme, means that you'll end up moving at some point which may be more than half way through the grinding. It's better in that light to stay and grind rock A even if you don't expect to get much additional data over the rock B grind. You use up some time and increase wear on your grinding tool.

With a fast communication delay, you can promptly instruct the rover to head over to rock B with consuming any time grinding rock A. The opportunity cost of changing your mind has been reduced.
« Last Edit: 07/24/2010 07:06 am by khallow »
Karl Hallowell

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #24 on: 07/24/2010 07:57 am »
That's a huge boost in power right there.
Err, 2x to be exact. So maybe 400 meters on a really good day instead of 200. Or a bit more since the non-driving loads would stay the same.

edit: actually, it should be a fair bit more, since a MER burns ~200 wh without driving at all. Even so, that gets maybe 3x current drive distance, assuming your 2x solar magically doesn't cost you anything.
Quote
That indicates to me remarkable power inefficiencies in the propulsion system for the vehicle.
You might want to think about that a little more. Top speed is not necessarily power limited, and driving distance isn't speed limited.

Since the vehicles *are* power limited, you'd think JPL would have tried to avoid "remarkable power inefficiencies"...
« Last Edit: 07/24/2010 08:09 pm by hop »

Offline khallow

  • Extreme Veteran
  • Full Member
  • ****
  • Posts: 1954
  • Liked: 8
  • Likes Given: 4
Re: A limiting factor in robotic planetary exploration
« Reply #25 on: 08/01/2010 10:03 am »
That indicates to me remarkable power inefficiencies in the propulsion system for the vehicle.
You might want to think about that a little more. Top speed is not necessarily power limited, and driving distance isn't speed limited.

Since the vehicles *are* power limited, you'd think JPL would have tried to avoid "remarkable power inefficiencies"...

The power consumption of the transportation system and the mass of the vehicle are known. As I pointed out, there should be enough power to go straight up at a speed well over the maximum speed of the vehicle on flat ground. I figure there was a tradeoff between efficiency and reliability and they decided to err greatly on the side of reliability. Or maybe there was a need to heat the wheels before use? I don't know. But it's reasonable to expect that there was more to the decision process than "Should we make the wheel motors efficient and not waste power?"
Karl Hallowell

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #26 on: 08/01/2010 09:27 pm »
The power consumption of the transportation system and the mass of the vehicle are known. As I pointed out, there should be enough power to go straight up at a speed well over the maximum speed of the vehicle on flat ground.
You are assuming top speed under normal conditions is power limited. Just because the vehicle *can* put 100 watts into the mobility system doesn't mean it does cruising at full speed on hard level ground.

Quote
But it's reasonable to expect that there was more to the decision process than "Should we make the wheel motors efficient and not waste power?"
When your conclusion implies gross stupidity on the part of people who otherwise seem quite smart, you might want to double check your assumptions.

Edit:
Some details of the MER mobility system: http://trs-new.jpl.nasa.gov/dspace/bitstream/2014/13269/1/01-2004.pdf
« Last Edit: 08/01/2010 09:51 pm by hop »

Offline khallow

  • Extreme Veteran
  • Full Member
  • ****
  • Posts: 1954
  • Liked: 8
  • Likes Given: 4
Re: A limiting factor in robotic planetary exploration
« Reply #27 on: 08/02/2010 03:33 am »
The power consumption of the transportation system and the mass of the vehicle are known. As I pointed out, there should be enough power to go straight up at a speed well over the maximum speed of the vehicle on flat ground.
You are assuming top speed under normal conditions is power limited. Just because the vehicle *can* put 100 watts into the mobility system doesn't mean it does cruising at full speed on hard level ground.

That was part of my argument. If the mobility system is not power limited, then what are the limiting factors? It's reasonable to suppose that they didn't not attempt to put in a faster system because of the communication delay.

Keep in mind the scenario above, if there's a few second communication delay, it becomes worthwhile to invest in a faster mobility system (perhaps heat dissipation being the biggest problem that needs to be fixed) even at the expense of instrument payload. You can cover more ground and investigate more things.

Edit: They're using a "harmonic drive" which has advantages of light weight and low volume, high gear ratio, and precise positioning.

Quote
Quote
But it's reasonable to expect that there was more to the decision process than "Should we make the wheel motors efficient and not waste power?"
When your conclusion implies gross stupidity on the part of people who otherwise seem quite smart, you might want to double check your assumptions.

My conclusion doesn't imply that. It implies (as I said later) that there was more to the decision process.
« Last Edit: 08/02/2010 03:46 am by khallow »
Karl Hallowell

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #28 on: 08/02/2010 04:15 am »
Quote
You are assuming top speed under normal conditions is power limited. Just because the vehicle *can* put 100 watts into the mobility system doesn't mean it does cruising at full speed on hard level ground.
That was part of my argument.
Huh ? your whole claim was that the MER mobility system must be inefficient because X speed times Y watts didn't add up. This is simply an error on your part. Do you understand why ?
Quote
If the mobility system is not power limited, then what are the limiting factors?
Motor RPM, which is driven by torque requirements, which is driven by mobility requirements, which don't depend on whether a human is driving or not. Plus there's no reason to go fast, because...
Quote
It's reasonable to suppose that they didn't not attempt to put in a faster system because of the communication delay.
Not reasonable at all. Total drive distance is power limited, not speed limited. Going faster would just run out your battery earlier in the day.
Quote
Keep in mind the scenario above, if there's a few second communication delay, it becomes worthwhile to invest in a faster mobility system
Why ? Total drive distance is still power limited.

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #29 on: 08/02/2010 02:43 pm »
Hop:  You're being too sensitive or something.  Karl is not implying "stupidity" on anyone's part.  My take is that he's asking questions like I do, but that he may not be using terms that pass your specification.  It's hard to understand what exactly it is that you're getting at.

Quote
Karl: There was no need for a high powered rover when the power couldn't be used for driving faster.

True, in the sense that the procedures for determining the rover's path were already known as checklists with time estimates designed to accomodate a range of known possibilities about the physical "facts on the ground".  Nothing can happen in less than twenty or so minutes, at least as directed from Earth.  So part of the rover's designed speed is to assure that it won't run into something or get into trouble in that time.  But there's also a huge list of the other factors.  Alternate power systems.  Mass and complexity of RTG's or solar panel.  Analysis of what could be done with 2x or 3x the power.  And on and on.

Within certain limits, there literally would be no advantage to greater power or speed, unless other mission objectives were also reconsidered.

Quote
Karl: There's no reason for scientists to hurry when the rover isn't going anywhere fast.

I'd say this is true.  You point out that they do have to hurry, since the time constraints of planning the MER cycles are very demanding.  Since the power consumption levels of the rover are known and its average ground speed abilities are known and its current environs are known; then it seems that the majority of the reasons for the time constraints is first, the determination of which objects are more likely to be of interest, then the determination of the rover's future path thru the objects encountered to the chosen destinations.

Thus there is no need for the scientists to hurry mainly because of the rover's speed.  They have to hurry for other reasons.  On that level, I'd say his statement is true.

Quote
That indicates to me remarkable power inefficiencies in the propulsion system for the vehicle.

My bold.  Then what should be the term used to explain what seems to be a huge difference between "theoretical" abilities and actual abilities?  There should be no emotional, political or intentional overtones accorded to his term "power inefficiencies". Certainly 2x or 3x the power boost would be "huge", at least in theory.

But Karl's first question was not ever answered, which is what gets me with all this agony over power consumption and speed.  What would be the effect of the time delay on the design of the rover?  And his second question wasn't answered either;  What explains the large difference between the theoretical and actual abilities?  Where does the inefficiency lie?  Emotionally, this is not to suggest that the inefficiency was designed into the system for reasons not related to the mission.

And his third question: what are the design tradeoffs?  Power, distance, speed, reliability?

Quote
Hop: Your whole claim was that the MER mobility system must be inefficient because X speed times Y watts didn't add up.

He didn't "claim" that there is inefficiency.  He asked, why does it look like inefficiency?  He isn't "claiming" that anybody is "stupid".

Since it seems clear that "Total drive distance is still power limited," which we all understood to be the case all along, how then, does communication time delay affect rover design?  Also, is there a breakdown of the rover's power budget?  Anyone?
Sometimes I just flat out don't get it.

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #30 on: 08/04/2010 03:27 am »
What explains the large difference between the theoretical and actual abilities
The assumptions underlying the "theoretical" abilities are wrong.

Now (as either of you would know if you had read the PDF I linked) the mobility system isn't spectacularly efficient. In ideal conditions, you can do quite a bit better, but the rovers don't operate under ideal conditions. Tele-operation wouldn't change the major constraints much: Mass ? No. Dimensions/packaging ? No. Mobility requirements ? Maybe relaxed a little, if you assume humans are less likely to get into trouble (dubious). Environmental conditions ? No. Reliability requirements ? No, unless you expect to drive further, in which case they go up.

Cutting out the computer power needed to drive autonomously might seem like a saving, but transmitting real time video and telemetry to some distant place would likely outweigh that by a wide margin.
Quote
And his third question: what are the design tradeoffs?  Power, distance, speed, reliability?
Power: You always want more. Duh ;)

Distance: Limited by power, for plausible solar (and probably RTG) vehicles, regardless of whether robotic or autonomous.

Speed: The current MER driving record is 220 meters. At the rated top speed of 5cm/s this would take less than an hour and a half (It takes a MER much longer, since it has to stop and think a lot, among other things.) To be speed limited in a 12 hour day, you'd need more than 8x the power available for driving. For a solar powered Mars rover, this appears completely unrealistic, so there's no real need to go faster than 5cm/s whether or not you are tele-operated.

As mentioned above, a MER doesn't actually average anywhere close to 5cm/sec, but that has more to do with their antiquated computers than tele-operation vs. robotic.

There would be some advantage to spending less of the day driving, even the total distance was the same. You could sit and do science longer (but that takes even more power), or have the rover awake for less time.

Reliability: See power.

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #31 on: 08/04/2010 03:56 pm »
Quote
The assumptions underlying the "theoretical" abilities are wrong ...[because in part]... the mobility system isn't spectacularly efficient
Well, there ya go, Karl.  And:
Quote
Power: You always want more. Duh.
What I get him saying is that the communications delay wouldn't affect the rover design all that much.  Which is one of the questions that we had.

Glad we got that cleared up.
« Last Edit: 08/05/2010 02:51 pm by JohnFornaro »
Sometimes I just flat out don't get it.

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #32 on: 08/05/2010 03:31 am »
Quote
The assumptions underlying the "theoretical" abilities are wrong ...[because]... the mobility system isn't spectacularly efficient
Um no. The "because" you inserted doesn't belong there. The inefficiency in the MER drive system (see PDF linked earlier, but note the efficiencies listed are for the harmonic system, not end to end) isn't sufficient to explain his claimed disparity.

"Not spectacularly efficient" does not mean "spectacularly inefficient".
Quote
Quote
Power: You always want more. Duh.
What I get him saying is that the communications delay wouldn't affect the rover design all that much.
IMHO, it would affect the rover design a lot, but not the basic power and mobility systems.

On Mars*, with solar or RTG, power will be maximized as much as practical in no matter what, because it's the ultimate limit of what you can do. Mobility is constrained by power and environment.

*Your mileage on other planets may vary.
« Last Edit: 08/05/2010 03:31 am by hop »

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #33 on: 08/05/2010 03:03 pm »
The "because" you inserted doesn't belong there.

So I edited my response to the more nit-pickingly correct, "because in part".

Quote
"Not spectacularly efficient" does not mean "spectacularly inefficient".

You are, as always, free to exercise your choice in drawing your conclusions based on your own thinking processes. 

Quote
IMHO, it would affect the rover design a lot, but not the basic power and mobility systems.

Thanks for that clarification, and thanks also for elaborating on what those differences might be.  I know, I know.  Do my own homework.

Quote
On Mars*, with solar or RTG, power will be maximized as much as practical...

I had no idea.

Quote
*Your mileage on other planets may vary.

I'll keep that in mind.
Sometimes I just flat out don't get it.

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #34 on: 08/07/2010 04:19 am »
and thanks also for elaborating on what those differences might be.  I know, I know.  Do my own homework.
Heh, never a bad idea, but not the intent. Post was long enough already.

It's a pretty open ended question. What's the mission goal, expected duration, where exactly is it operated from etc.

If the mission and mass budget were similar to MER, I think the vehicle could end up pretty similar (whether there is any plausible scenario where this would make sense is a different question...) You'd need better communications gear and cameras, but technological advances will cover some of that. As I mentioned earlier, the advantage would be operational efficiency: Instead of waiting a whole sol to find out the IDD didn't go in the right place, you would adjust it and carry on. Similarly, driving would be much more efficient, since you wouldn't have to plan whole drive from one vantage point, or rely on very slow, inefficient auto-nav.

Lunokhod is another example, but power is much easier on the moon, and computer technology has advanced a bit from what was available to the Soviets in the '70s

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #35 on: 08/07/2010 02:16 pm »
It's hard to do one's own homework when one doesn't know what the assignment is, and when the student's questions are criticized as being inadequate.

Skip down to the ******** line below if you wish:

Sometimes a recap is helpful to all the students, particularly those who might be shy and don't ask questions.  From the OP:
Quote
In five years of exploration ... these two small rovers have driven no more then a couple of miles. Why? In my opinion the answer is that the communication delay with Earth... doesn’t allow for a real time piloting of these machines.
And the simple answer:
Quote
Power is the limiting factor, not communications delay.
Sivodave rephrases his question:
Quote
Well but i think that the communication delay poses some kind of limitation for a robotic exploration mission. maybe is not a big issue but relevant somehow.
Hop's obvious answer also adds the important bit that, in my phrasing, analysis time far exceeds light speed of communication, so that distance, while a factor, isn't a fundamental factor:
Quote
Of course it's a factor. ... MER's communication round trip time is much longer than light time....
and gives a brief summary.  Whoa! Simonbp offers an alternate view:
Quote
Nope. It's not power. It's the darn scientists.  Even at their best, the rovers only drove at ~20% of their theoretical speed...
But here, I think he's talking about time in general and not making a suggestion that the rover should be "maxing out the odometer". Several subsequent posters emphasize power, power, power.  Good.  We get that.

Garret offers another take:
Quote
A planetary scientist would probably be just as happy to send a static probe...
but this is just an interesting sidetrack.

Karl asks the question that I am interested in also:
Quote
Would these rovers have been designed the same way, if the communication delay were a mere 2.5 seconds round trip?
Hop gives the obvious answer; it would be a "different beast".  Okaaay. 

Then more discussion about power and scientific scheduling, which is fine and helpful.  Then Hop makes some kind of error, what I sometimes call jumping to grand conclusions on one data point:

Karl: "That was part of my argument..."
Hop: "Huh? Your whole claim was..."

My psychoanalysis here suggests that whatever thought process which took "part of an argument" and fashioned it into a "whole claim" is similar to the thought process that previously found implied stupidity where there was none.  This type of mistake derails conversations, and impedes the flow of information between "enquiring minds", who "want to know".  Continuing statements like; "as either of you would know if you had read the PDF I linked" cloud the discussion even further.  I read that PDF.  It doesn't address the comparison of communications delays, the questin at hand.  I think there are too many details demanded of the questioner, and that the spirit of the question gets lost in these details.

Look at all that verbiage above, just to recap.

***************

Quote
Heh, never a bad idea, but not the intent. (even tho that was the question) Post (Reply # 32, I assume?) was long enough already.

The question again, worded slightly differently:

Changing one parameter about MER, the communication delay, from 20 minutes to 2.5 seconds, what would the ramifications on the rover design be?

One possible, partial answer, which doesn't have to address all or any of the other possibilities:

Quote
I think the vehicles could end up pretty similar

So really, all things except communication delay being equal, what the problem is with the OP?

Quote
the communication delay with Earth (20 minutes for sending a message, other 20 for receiving the answer) doesn’t allow for a real time piloting of these machines.

It's not really about power or bandwidth, is it?  Unless power or bandwidth is what you're looking at, instead of just communication delay.  Power and bandwidth are mission and rover specific, and cannot compensate for distance.  AI, if it develops its abilities over time, can ameliorate local terrain accidents, but will not change the communication time except that time will not be lost to some types of accidents.  The time of science can not be predicted at all and is completely independent of communication delays, because we are boldly going where no rover has gone before.  In some ways, we can take advantage of time delay for creating sub-mission profiles, as long as the rover isn't incapacitated by a terrain accident.

If we had an orbiting manned martian eco-lab (with attendant hotel, but I digress) we could actually get more done with the rovers we have, it seems, if one allows the purely hypothetical example.  We could drive faster and further over clear and uninteresting terrain.  We could quickly avoid obstacles that exceed AI capabilities.  We could abandon false leads faster.  We might be tempted to push the rover beyond its specifications, however and we could still run into something at these breakneck speeds.  All in all, communications delay is a, which is not to say the only, legitimate "limiting factor in robotic planetary exploration", as Davide wonders.

Clearly, the ability to pilot these machines in real time would warrant a different analysis of power, mobility, and other rover characteristics.

Two points for Davide.  You're in Gryffindor, right?
Sometimes I just flat out don't get it.

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #36 on: 08/07/2010 10:53 pm »
All in all, communications delay is a, which is not to say the only, legitimate "limiting factor in robotic planetary exploration", as Davide wonders.
No, read the passage you quoted from the OP again.

Quote
In five years of exploration of the martian soil, these two small rovers have driven no more then a couple of miles. Why? In my opinion the answer is that the communication delay with Earth (20 minutes for sending a message, other 20 for receiving the answer) doesn’t allow for a real time piloting of these machines.
(emphasis added)
According to him, communications delay is the reason.

This is simply wrong*. The fundamental limit is power, with communications bandwidth next. (Assuming launch mass and budget are constant, otherwise those are obviously first.)

Real time control would give you a decent improvement in operational efficiency. Maybe you could get 2x or 3x done in the same amount of time. The distance would still be measure in miles per year, and there are other ways you could get the same gain.

Having continuous high bandwidth communication coverage of MER (along with sufficient ground personal to take advantage of it) would give you almost the same improvement in operational efficiency, even with the 20-40 minute round trip time. Instead of using autonav, you do direct drives all the time, planning the next one as new terrain came into view. The total distance you could cover and efficiency of doing so would be nearly the same as driving in real time. Most of the sols lost due to sequences not working correctly would also be eliminated.

A 2x increase in power would also give you a similar improvement, or let you do things that are simply not possible on the current vehicles.

So to identify the time delay as the reason the MERs have only gone as far as they have is simply wrong.
Quote
Clearly, the ability to pilot these machines in real time would warrant a different analysis of power
Not really. For power the results would be about the same. Whether you are operated in real time or not, it is the fundamental limit on how much you can do, and is ultimately limited by mass, money, and the target environment. For a given amount of money, you are going to get the same power in a given package. For a MER type vehicle, you will be power limited, so in either case you get as much as you can in your budget.
Quote
mobility
Nope. This is limited by power (which is fixed, per above) and mission requirement for what kinds of terrain you want to be able to deal with. It won't change much whether you are tele-operated or not. As I showed in an early post, you don't need to go any faster for any realistic amount of power.

* It's also wrong in another way. Opportunity just crossed the 22km mark, which is a bit more than a couple miles ;)

Offline Hungry4info3

  • Full Member
  • ****
  • Posts: 525
  • Liked: 163
  • Likes Given: 81
Re: A limiting factor in robotic planetary exploration
« Reply #37 on: 08/08/2010 12:10 am »
Why does the fact that power is the limiting factor to how far the MER rovers can go keep having to be reiterated?  :-\

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #38 on: 08/08/2010 03:26 pm »
Quote
Real time control would give you a decent improvement in operational efficiency. Maybe you could get 2x or 3x done in the same amount of time.
And so, in Davide's small statement, communication delay alone, not changing any other parameter, is "the" limiting factor in this single example, especially given his stated concern with "real time piloting".  And as I said, sorta implying that there are many other possible missions, factors and whatnot, it is "a", but not "the" limiting factor.  Now maybe his intent is indeed to look to the larger issue, as is implied by the thread title.  Which is fine also. 

I have not learned a single thing in this thread so far.  If I had been the first poster, this is what I would have said:

<<
As you are probably aware, power consumption of any rover limits its abilities; there's only so much that you can get; it can only be but so efficient; and you have to both purchase and successfully launch the rover.  I think everyone would agree that "real time piloting" would be the best situation hands down.  Navigational problems could be dealt with in real time, and prompt decisions could be made based on the interest level of ground features as they came into view.  We could have a quick look around, and then spend 40 minutes driving straight to what looked like an Egyptian obelisk, for example.

Clearly, a casual analysis would conclude that a manned orbital martian laboratory would cost a lot more than the the comparably equipped terrestrial lab operating the rover with the 40 minute delay, so that's what we're stuck with for the moment.  And of course, the time delay problem gets even worse the farther away the rover gets.  Like Pluto?

The related question that occurs to me, as I read your post is:  How would a real time rover be different than a time delay rover?  To simplify the problem, I would remove terms from the equation of the question.  So, assuming the exact same mission, what would be changed in the rover's physical manifestation, and in the related human procedures for its operation and mission evolution? 

The first things that come to mind are power, efficiency and reliability of drive train, mobility, communications gear, addditional communications sats, mission planning procedures.... what else?
>>

But no.  Instead we get: "it would affect the rover design a lot" and "I think the vehicle could end up pretty similar" ending up with "Not really".  Confused?  This is what analysis paralysis looks like, you lurking NASA watchers.

{quote]Why does the fact that power is the limiting factor to how far the MER rovers can go keep having to be reiterated?{/quote]

I'll bite.  Perhaps it is because, in part, that, the error of a poster characterizing 23km as "a couple of miles" is so wrong as to merit no further discussion.  Nobody has chosen to briefly and concisely educate us on a comparison of rover design on the basis of communications time delay. 

Clearly, the rovert's (Named after Stephen, of course) science productivity can go up quite substantially on this issue alone, as Davide correctly surmises, all else being equal.  Determining this productivity can provide important insight into the value of a martian orbiting lab, which promises to be an necessarily expensive potential NASA mission, especially as the issue of an ecosystem up there becomes closer to becoming accepted public knowledge.

Here's some homework:
http://www.eng.buffalo.edu/mechatronics/publications/conference/PatLeeKrovi_OutputSynchronizationForTeleoperationOfWheelMobileRobot.pdf
Ouch.  Not bad tho, eh?  I gotta read this one.

Haptic interfaces are a related field of inquiry:
http://deepblue.lib.umich.edu/bitstream/2027.42/60859/1/paulgrif_1.pdf

For $50 (used), you can read about it:
http://www.amazon.com/Teleoperation-Robotics-Progress-Astronautics-Aeronautics/dp/1563470950

Chapter Three has a section called: "Coping with Time Delay".  I just bought a copy, FWIW.

Another intersting title:
http://www.eng.fiu.edu/mme/robotics/elib/MED04%20Kusadasi%20Turkey%20-%20Effects%20of%20Time%20Delay%20on%20FF%20Teleoperation.pdf

Here's one of some limited applicability, titled: "Autonomous Rovers for Mars Exploration"
http://74.125.155.132/scholar?q=cache:wJ0SJLTRtAMJ:scholar.google.com/+teleoperation+of+rovers&hl=en&as_sdt=80000000000000&as_vis=1

More about time delay:
http://www.scielo.org.ar/pdf/laar/v33n4/v33n4a02.pdf

Personally, I'm getting overwhelmed with info.  Here's one more:
http://robotics.estec.esa.int/i-SAIRAS/isairas1999/s14-01.pdf

Oh my garnet.  Remember, YMMV on other planets...
Sometimes I just flat out don't get it.

Offline TyMoore

  • Veteran
  • Full Member
  • ****
  • Posts: 478
  • Eureka, CA, USA
  • Liked: 3
  • Likes Given: 1
Re: A limiting factor in robotic planetary exploration
« Reply #39 on: 08/08/2010 04:50 pm »
Just tossing in my (I'm-not-an-engineer) two cents worth:

The 'low' efficiency of the MER drive system has more to do with the 'interaction' of several related factors: the rather large gear reduction ratios in the planetary gear boxes, the low Martian surface temperatures, and the viscosity of the Castrol Braycote Micronic 601EF vacuum grease plating those gears.

I think the correct answer to all of this is this: the operating efficiency was 'good' enough to achieve the mission requirement, which was a duration of 30 days. Almost eight years later, we are seeing the rovers literally wearing themselves out...it is a wonder they lasted this long at all!

As a side: it sure would be interesting to take a look inside the various components of the rovers to see how far the Martian dust-fines have migrated past various seals inside the various mechanisms of the rovers.

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #40 on: 08/08/2010 05:12 pm »
I saw that thing about the grease also, and made a mental note for use on a lunar lander for the ice craters.  Really cold.  Bound to be dust.
Sometimes I just flat out don't get it.

Offline hop

  • Senior Member
  • *****
  • Posts: 3352
  • Liked: 553
  • Likes Given: 891
Re: A limiting factor in robotic planetary exploration
« Reply #41 on: 08/08/2010 09:27 pm »
But no.  Instead we get: "it would affect the rover design a lot" and "I think the vehicle could end up pretty similar" ending up with "Not really".  Confused?  This is what analysis paralysis looks like, you lurking NASA watchers.
No, this is what off the cuff observations of a semi-educated amateur observer look like! Just to be clear: I do not, and have never worked on space missions of any kind. What I post here is my on opinion.

Some of your confusion appears to stem from not reading or understanding the things the quotes refer to: The "not really" does not contradict "vehicles could end up similar", and the reasons for it are explained.

As I said before the answer you get depends so much on the assumption that you can come up pretty much anything.

A notional "imagine Earth was only 2.5 light seconds away from Mars" MER could be very similar, but that still depends on the assumptions. How much communication bandwidth is available, how often and how much mass and power does it cost ?

If you assume the same communication bandwidth, availability and power cost as a real MER, it simply doesn't work. 128 kbit/sec for <10 minute intervals a few times a day isn't enough for useful teleoperation. 

If you assume constant, high bandwidth communication for the same power cost it will work a lot better, but as I pointed out previously this would be a big gain in operational efficiency even if light delay stayed the same.

The obvious conclusion is that bandwidth and communications availability are bigger limits on MER than light time.

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #42 on: 08/09/2010 02:01 pm »
If they are truly off-the cuff observations, the signal to noise ratio can be increased by choosing to minimize comments like, "if you had read the PDF's..."  Another hint that I would offer would be to de-emphasize spinning, "does not contradict" and "depends on assumption", and emphasize possible answers, including, as always, alternate assumptions for consideration.

The OP is more properly interpreted as assuming only the change in comm time.  The addition of other possible changes in terms does not further the answering of that one question.

Again, there is no pressing need, other than individual choice, to add "if you assume" to the OP.  Certainly, they are not unreasonable assumptions to postulate.  It is the admonitions which are unreasonable and which detract, and it is the frenzied sequential justifications which confuse and dispirit other amateurs.  Lastly, in this case, are the simply incorrect statements like "not reading", which are inconsistent with the quiet assumption of learning that at least I grant other amateurs.  I, for one, read very carefully; nevertheless, this is no guaranttee of total understanding, but I do not confuse the two concepts.

Part of my personal quest and interest could be summarized thus:  Primitive man builds a martian rover.  Primitive man knows that martian rovers have been built before.  Primitive man's brain is virtually indistinguishable from the brains of those who have built those rovers, even if his personal experience and experiences differ from that of the builders by a wide margin.  Primitive man notices that size of checkbook appears to be the only salient difference in this case, but chooses to continue his quest, largely for reasons which escape the understanding of the fundamental scientificist.  FWIW and YMMV, as usual.

Part of my strategy in understanding these complexities, is to simplify the equations.  If we simplify one factor in this MER scenario, comm time, how would that affect rover design?  Obviously, an orbiting lab, cuts comm time from the equation.  An orbiting lab, kemosabe, is something that we (which is me and at least one other person, who I know) would like to see up there at Mars.  It would be a type of flexible path mission which could determine, and then possibly work to preserve, an ecosystem.

There's a ton of reading material for me to assimilate.  I'll share some of that later.
Sometimes I just flat out don't get it.

Offline JohnFornaro

  • Not an expert
  • Senior Member
  • *****
  • Posts: 10974
  • Delta-t is an important metric.
  • Planet Eaarth
    • Design / Program Associates
  • Liked: 1257
  • Likes Given: 724
Re: A limiting factor in robotic planetary exploration
« Reply #43 on: 08/10/2010 05:26 pm »
First off, this report: "OUTPUT SYNCHRONIZATION FOR TELEOPERATION OF WHEEL MOBILE ROBOT"  made my eyeballs roll back so far in my head with math that I actually had to bang my head on the wall to get them to come back in focus.

This one: "Effects of Time Delay on Force-Feedback Teleoperation Systems" was easier to follow.

Teleoperation, briefly described:

Quote
Teleoperation describes two systems that are distant from each other and coupled in a way that both send and receive commands from each other. The information sent from the master to the slave controller is the position and/or velocity command and the information sent from the slave to the master is usually the force command. The force feedback from the slave provides valuable information to the master to get the feeling of the conditions the slave faces in order to improve the operator’s ability to perform the manipulation with small errors, which could save the slave from exerting unnecessary amounts of force on the environment.

And the issue of interest, regarding martian rovers:
Quote
The time-delay between the master controller and the slave robot on the communication lines has arisen as a dominant factor of instability in teleoperation.

Now this is in 2004, and I don't know how much things have improved since then.  But a haptic interface, which is where there is "touchy-feely" feedback from the rover to the human manipulator, is readily understood as improving the man-machine interface.  If you know how hard to press against something, you can conserve energy by not pressing too hard.  If a gentler or harder touch is needed, say, to get out of a stuck situation, then a haptic interface also is quite helpful.  My understanding of MER is that there is not much of a haptic interface.  This report is only about the MatLab simulation of a one Degree of Freedom setup, but a haptic interface is still a good idea in general, I'd say.  The report  suggests a "wave variable technique" as a possible means of accommodating time delay, and recommends the usual: "The next step for the continuing study of wave variables will be the investigation of the effect of..."

One of the good things about this report is that it educates the reader a bit about the building blocks of a teleoperated robotic system.  There's some basic discussion of "Joint Actuators", and "Joint Sensors", and other pieces of the robotic simulation.

It should be obvious that adding such an ability is to change the ability of the rover, but that is indeed my question:  What are the changes to rover design based on the comm delay?

Paul Griffiths' dissertion on haptics is also interesting, long, and more focused on medical solutions, but he presents a slightly different description of a haptic interface:
Quote
haptic device is typically a hand-held interface, such as a handwheel or pen-shaped stylus. By manipulating the device, the user navigates through a computer rendered virtual environment, and motors in the device push against the user to synthesize a mechanical interaction with the virtual environment.

He mentions the idea of "haptic rendering", which I take to mean the conversion of a computer depiction of force from the force felt by the robot to something that the human hand/brain system can readily interpret and respond to.  He talks about "impedance type" devices, which is where, I think, the previous paper leaves off.  He also notices that "stability in the face of significant delay between the master and slave requires special compensation which necessarily degrades the accurate rendering of the remote environment dynamics".  Since we'll eventually need long distance medical attention, say when a martian explorer gets a splinter, this is related to our rover questioning somewhat.

Personally, it's hard to stay focussed because there's so much to learn: "Although feedback may be used to attenuate sensitivity, feedback may also amplify sensitivity."  It's possible to have too much of a good thing!  Anyhow, becausee this paper is mostly medical, I only read to p. 11 out of 111.

This paper: "ON-LINE ESTIMATION OF COMMUNICATION TIME DELAY IN A ROBOTIC TELEOPERATION SYSTEM" hammers home the thought of the OP: "In recent years, several experiments have proved that the execution time needed by a human operator to do a task with a robotic teleoperation system, is significantly dependent on the communication time delay..."  In its introduction, it goes on to describe the typical human strategy, followed today by the MER team: "move and wait".  In other words, we have to have a series of discrete reference commands, because of time delays.  For the last twenty years or more, "much effort has been put on designing control structures capable of compensating communication time delays..." 

Since variable time delays are worse than known ones, this paper addresses the estimation of those delays, again using lotsa greek letters and involved flow charts which defy thorough understanding from a casual read, which is a judgement on the reader, as we all know.  I do like the "forgetting factor k", and plan on using it often. 

This brief document survey has convinced me that the limiting factor in robotic planetary exploration is indeed the time delay, as the original poster intuits. 

One of the benefits to real time operation is the addition of a haptic interface, which might suggest a different approach to the problem of mobility.  In addition, there should be power savings due to the haptic features such as quitting a project that is too hard to do.  There whould also be power savings due to decreased transportation times to interesting sites, since some decision making processes could be shortened dramatically.  Thus, more science/watt might be expected.
Sometimes I just flat out don't get it.

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
0