Getting Unstuck in West Valley Options

[tty]

To return to nprevs suggestion about rocking, this is the preferred way of getting yourself loose when stuck in snow here in Sweden: You gun the car for a moment, then unclutch and let it fall back. However it will move a little bit past the equilibrum position, so you catch it on the return and gun it again and so on. If you time it right the oscillation will gradually build until the car gets loose. I was taught the technique for snow, but it works just as well for desert sand and laterite clay (I've tried both). Surprisingly the technique does not seem to be generally known in other parts of the World.
I don't know if such short bursts of power are practical for MER, and the timing as I said, must be precise. Possibly it could be tried in the test-bed.

Also I would like to comment on the possible use of the IDD for pushing. It might make a difference. I remember a case when two people pushing (I was one of them) was enough to get a ten-ton truck unstuck.

[Fran Ontanaya]

She could also wave the arm and become a Tesla's oscillator.

[stevesliva]

Surprisingly the technique does not seem to be generally known in other parts of the World.

Too many automatic transmissions here in the states, although I was familiar with the technique with my old econobox that also didn't have power steering. It was the best possible car for detecting slippery road conditions-- both the clutch and the steering wheel let you know if you were losing traction.

I don't think the rovers have a clutch of either variety, though.

[nprev]

I learned to drive on an old standard transmission truck in the mountains of Montana & lived in Alaska for a few years, so have had to rock out of a few slippery situations...

Problem is that doing this old-school is very dependent on real-time feedback to the driver; we don't have that luxury with the MERs. If this is tried, I'm hoping that the major effect is to precess the fore end of Spirit downslope (to the left) & let gravity help with the final extraction.

[Guest_Oersted_*]

I did that rocking back-and-forth thing quite often, but I don't know what the clutch needs to be used for? I just had the van in first gear and gave it gas everytime I started on the forward motion. Basically it is like pushing a swing, you give it an extra push everytime you reach the highest point on the backswing. It is also how Robert Langdon gets out of an iffy situation involving a big book cabinet and an armed glass window in the new Angels and Demons movie

That technique really needs timing, so I think it must be very difficult to perform on Mars.

[hendric]

Well, I don't know about timing, but it should be possible to use the IMU's accelerometers to measure any rocking and apply throttle appropriately. You push in the clutch normally on the backwards roll. Unfortunately, I don't think the wheels turn anywhere near fast enough to initiate the rocking to begin with. What's more likely is that she'll have to "swim" out of the sand by sheer brute force, using the treads as paddles to move sand from the from in front to behind the wheels. The problem is that it is a race between getting out, and the wheels sinking below the top of the hubs or high-centering on the bottom. One experiment to try is figuring out which part of the sol gives us the best traction, similar to Nomadd22's idea of waiting until winter. I'll bet that the consistency of the sand changes with different times of the day/night, as the relative humidity increases the sand/dust grains will stick together better. Plus, the area beneath Spirit is probably the coldest place on the Columbia Hills, so it'll collect any moisture around, a la freezer burn. So sitting and thinking for a bit is a good thing.

[nprev]

Yeah, like I said, wasn't really envisioning rocking in the sense of getting out of a snowdrift. Getting the nose turned downslope via precession is about the best we can hope for IMO.

Still, though, it would be interesting (and useful!) if the MERs could execute a complex autonomous subroutine as you described, Hendric. The required response-to-feedback time is pretty short.

[RoverDriver]

Something was commanded (site/drive numbers have changed) but I can't see any movement.

This was a short duration command on the LM wheel to diagnose the wheel stall we had on 1899. I haven't looked at the data yet, so I am not sure of the results.

Paolo

[RoverDriver]

Yeah, like I said, wasn't really envisioning rocking in the sense of getting out of a snowdrift. Getting the nose turned downslope via precession is about the best we can hope for IMO.

Still, though, it would be interesting (and useful!) if the MERs could execute a complex autonomous subroutine as you described, Hendric. The required response-to-feedback time is pretty short.

Unfortunately, the interval between commands is one second. That limits the response time. In addition, sensors are queried every 1/8 sec. Way too slow for any real-time control of the vehicle. Some serious coding would be needed, tested on the ground, uploaded on the vehicle and verified on Mars before we would be even attempt to use it.

Paolo

[Astro0]

Alan: Some wheel movement thisol?

According to their filenames, both of those images indicate that they were taken on Sol 1897. So this was preseumably some of the early moves after getting stuck. The images just came down on the later Sols.

[ngunn]

Some serious coding would be needed

FWIW here's how it looks to a complete outsider in these matters:

Elegant idea (using resonance as an amplifier) + proven effectiveness on Earth + outwith present capability + possibility of developing software to do it + potential benefit to other rovers, present and future + serendipitous availability of 0.4g sandbox = highly worthwhile project.

[djellison]

The rocking technique ( I've used it as well ) involves free-wheeling back to the bottom of the little trench you dig, then throttling back up. MER can't free wheel. Essentially, it would just be driving backwards and forwards in little steps. To me- that sounds like an excellent means of digging the wheels in further, without making any progress.

[Astro0]

Hopefully we'll get some better visualisations soon, but I thought I'd play with a Rover model, some flour (for the soft material) and topped with some cinnamon (for contrast) to simulate for me what's going on. I created a ridge structure under the 'soils' to create a spot where loose material is stacked up against some kind of ridge, as suggested by some of the images.



I've tried to get the wheels in at roughly the right depth and haven't tried putting rocks underneath.
Just looking at it though, there's not much room between ground and WEB.
The rocks themselves though don't look that big. I guess it's more a matter of how far the RL/R and ML/R wheels have dug in.
What we really need is a look underneath Spirit.

[SpaceListener]

Problem is that doing this old-school is very dependent on real-time feedback to the driver; we don't have that luxury with the MERs. If this is tried, I'm hoping that the major effect is to precess the fore end of Spirit downslope (to the left) & let gravity help with the final extraction.

I thought it too but I am still uneasy since this is a like a chess game. By taking the advantage of the left side using the gravity help but after, toward the Braun hill (south), there are small slope that Spirit must overcome that, I think it would be very tough. In order to evaluate the north evacuation way, it is necessary to study about how is the soil surface toward the north.

[RoverDriver]

Hopefully we'll get some better visualisations soon, but I thought I'd play with a Rover model, some flour (for the soft material) and topped with some cinnamon (for contrast) to simulate for me what's going on. I created a ridge structure under the 'soils' to create a spot where loose material is stacked up against some kind of ridge, as suggested by some of the images.
...

That is pretty awesome! If only that model had working rocker-bogie system the state of the wheels would have been represented with even more fidelity. If only I could use flour in the testbed but I was told it is a fire hazard. And cinnamon too! The ISIL would smell so good!

Paolo