“Crawl Control” for the Cybertruck?

[Jhodgesatmb]

Crawl mode has been discussed on the forum before and it would indeed be odd if Tesla doesn't implement it. Different people have different notions of what crawl mode means .

That said, perhaps you should write to Tesla and describe this to them. 'We' cannot implement anything.

Sponsored

 

[slomobile]

Crawl mode has been discussed on the forum before and it would indeed be odd if Tesla doesn't implement it. Different people have different notions of what crawl mode means .

That said, perhaps you should write to Tesla and describe this to them. 'We' cannot implement anything.

If I were to write Tesla, it would be to suggest scriptable low speed control algorithms.
So that each of us could write the control algorithm we want for our individual use. Submit the script to Tesla engineers for first round safety evaluation, and schedule a time for Tesla to upload the feature to our individual truck while they monitor. We operate the script under their supervision.
Tesla makes the call to allow us to continue unsupervised or not. In exchange for Tesla oversight (which costs money) they get a semi exclusive license to the code we created. And the right to distribute our personal drive mode to other Tesla customers.

 

[JBee]

I consider this a must have feature. I've not seen how it is done by Toyota, but now I will have to test drive theirs to see.

I'm reminded of Grandpas tractor. Pick a slow gear. Set your constant ground speed with the hand throttle quadrant, and control steering with left and right dig brakes.

In CT crawl mode, maybe foot brake and throttle can be used to bump down/up the setpoint of a very low speed cruise control. With foot brake continuing to operate hydraulic brakes also. Yoke steering angle could simultaneously control front wheel direction and wheel speed per side for half track style steering. Left foot braking while applying throttle would increase rate of turn without changing the base crawl rate. Holding the inside wheel while speeding up the outside wheel. Center touchscreen could display the current base crawl rate and allow it to be adjusted with finger slide for those moments when feet are too busy and operation gets confusing.

Although your description and control methods sound neat they are somewhat unnecessary.

Crawl mode would only have to remap the throttle pedal so that you have to use more pedal travel for smaller speed changes. A crawl speed can be set using the cruise speed controls.

If you have a quad motor CT you end up with something called torque vectoring, that uses the wheel speed sensors and independent wheel motors to maintain wheel rpm to match the ground speed and to create an electronic differential that uses the steering wheel angle sensor to modulate wheel rpm between left and right and front and rear so that each wheel is precisely matched to the ground speed of each wheel.

What this means is that every wheel is always spining at the right speed already to to match the ground, if you turn the steering it automatically changes the wheel rpm to facilitate the turn. It in fact has to do this because there are no open differentials on a quad motor CT, and if it didn't it would be like driving around with two locked differentials and you would drift everywhere like on ice.

So with a QM and torque vectoring the whole "skid steer" like a tank is always and already active and changing wheel speeds to keep the car stable and on the road.

Another way of thinking of it is like a "always on" ESP on standard cars, but that it is actively not allowing wheel slip. All this is done behind the scenes without the driver having to do anything for it.

Anyway, its going to be pretty epic offroad with torque vectoring.

 

[slomobile]

What this means is that every wheel is always spining at the right speed already to to match the ground, if you turn the steering it automatically changes the wheel rpm to facilitate the turn.

That is a nice theory on flat ground with traction. Skid steer implies some skidding, not optimum rolling. For that reason, I cant even say my idea is good enough to defend. I haven't put enough thought into it.

Sometimes a bit of wheelslip is necessary because none of the wheels have sufficient traction and torque vectoring means which wheel is churning the mud more effectively. I have yet to see a sensor that can figure that out. Maybe a yaw sensor.

But in off camber situations, you don't want your vehicle to try to outsmart you. You definitely don't want it making sudden accelerations because a bump caused you to jerk the wheel. My idea was bad for rock crawling, might be fun for mudding.

Even if a computed method may be more theoretically efficient, when crawling, you want the vehicle to behave consistently. Predictably. Even if that is not optimally. Quite often the direction the steering wheel points is not the direction you want to go. Is CT smart enough to choose the right line?

 

[Crissa]

Touch screen is bad for any controls that are needed for when you're jostled by rocks.

(That's my experience anyhow)

-Crissa

 

[slomobile]

Touch screen is bad for any controls that are needed for when you're jostled by rocks.

(That's my experience anyhow)

-Crissa

Agreed. That is why touchscreen is only a secondary control. Accelerator and brake are primary.

I think people would understand if they could drive an old tractor, or an old car with throttle quadrants. A set and forget hand throttle seems absurd to us that are so accustomed to a throttle always returning to zero. But it was not always that way. Driving such a beast really messes with your mind for a few minutes, then it feels completely intuitive. We (humans) have a remarkable ability to adapt to, and compensate for machines.

 

[ajdelange]

A set and forget hand throttle seems absurd to us that are so accustomed to a throttle always returning to zero. But it was not always that way.

Airplanes? Boats?

 

[JBee]

That is a nice theory on flat ground with traction. Skid steer implies some skidding, not optimum rolling. For that reason, I cant even say my idea is good enough to defend. I haven't put enough thought into it.

Sometimes a bit of wheelslip is necessary because none of the wheels have sufficient traction and torque vectoring means which wheel is churning the mud more effectively. I have yet to see a sensor that can figure that out. Maybe a yaw sensor.

But in off camber situations, you don't want your vehicle to try to outsmart you. You definitely don't want it making sudden accelerations because a bump caused you to jerk the wheel. My idea was bad for rock crawling, might be fun for mudding.

Even if a computed method may be more theoretically efficient, when crawling, you want the vehicle to behave consistently. Predictably. Even if that is not optimally. Quite often the direction the steering wheel points is not the direction you want to go. Is CT smart enough to choose the right line?

The beauty of 4 independant motors with vectoring is that it doesn't matter what type of ground you are transversing. All it needs to do is match wheel speed to ground speed, whilst the acelerator controls velocity gain and steering heading change. If you want more wheel spin accelerate more like on any other car, if you want to tank turn, turn wheel and it will instantly increase outside wheel rpm and decrease inside corner rpm. They already have 3 axis sensors for ESP etc.

And if you can't steer in the direction you want to go, what other control method do you use to achieve that on a car?

 

[charliemagpie]

I ordered with FSD … One day, 'FULL' will mean Full

 

[HaulingAss]

If I were to write Tesla, it would be to suggest scriptable low speed control algorithms.
So that each of us could write the control algorithm we want for our individual use. Submit the script to Tesla engineers for first round safety evaluation, and schedule a time for Tesla to upload the feature to our individual truck while they monitor. We operate the script under their supervision.
Tesla makes the call to allow us to continue unsupervised or not. In exchange for Tesla oversight (which costs money) they get a semi exclusive license to the code we created. And the right to distribute our personal drive mode to other Tesla customers.

Throttle control in an EV is child's play. A mode for extreme offroad will be very useful but I don't see any point in different users needing different settings beyond that because the nature of an electric drivetrain is for it to be telepathic in term of direct feel and controllability. All that's needed is a remapping of the throttle to make it much less sensitive at very slow speeds.

And, yes, I have been driving offroad vehicles for decades in all kinds of environments so I know what is needed.

 

[rr6013]

Throttle control in an EV is child's play. A mode for extreme offroad will be very useful but I don't see any point in different users needing different settings beyond that because the nature of an electric drivetrain is for it to be telepathic in term of direct feel and controllability. All that's needed is a remapping of the throttle to make it much less sensitive at very slow speeds.

And, yes, I have been driving offroad vehicles for decades in all kinds of environments so I know what is needed.

Agreed. Your right foot is well educated.

Selectable profiles are on my GIMMElist.
STEER bias’ power 60/40 to front wheels steering input integrated

BALANCE fine tune power distribution

MAP selector for motor Torque profiles

REGEN motor braking strength

 

[slomobile]

Throttle control in an EV is child's play. A mode for extreme offroad will be very useful but I don't see any point in different users needing different settings beyond that because the nature of an electric drivetrain is for it to be telepathic in term of direct feel and controllability. All that's needed is a remapping of the throttle to make it much less sensitive at very slow speeds.

And, yes, I have been driving offroad vehicles for decades in all kinds of environments so I know what is needed.

Have you driven offroad by independently and finely controlling 4 independent wheel velocities and steering angles?

Some wheels forward, some reverse, and 4 independent steering angles? (if we are lucky)
A few people in the world have done that in full size vehicles. Maybe you are one of them, but it is far from common knowledge.
I have done it with 4 wheel drive robots I've built. I've learned that there are many bad ways to do it and many good ways that depend greatly on the situation you find yourself in. I am highly skeptical of anyone that claims to know the one best way and calls it simple. To be clear, these are only small gains over the typical methods. But they are gains.

I've written thousands of lines of code doing this and the hardest part has been the user interface. Frankly I'm pretty bad at it. There are just too many options. Simplifying the user interface by limiting your options is one very effective way. It is not the only way, but may be best for the masses.

A drivetrain with so many options only feels telepathic after a large amount of testing, tuning, code, repeat, and somewhat predictable environments.

You know how to drive offroad roughly the same way you always have. There have been innovations, like rear steer, hydraulic steer, hand dig brakes, shock improvements, suspension geometry improvments... We have an opportunity to discover entirely new ways to drive. We should record and share them.

Some examples:
Steer rear wheels far left to get front of left rear tire just barely outside a rock obstacle. Forward a couple inches, steer rear wheels right to pull the entire rear of the vehicle left, onto a new line.

Steer rear wheels outward, front wheels inward, left wheels forward, right wheels reverse, to do a right turn in place.

On brake, steer all wheels inward for better hill holding control downhill. Upon throttle, while yoke turned left, only steer left front wheel outward and allow its velocity to be dependent(freewheeling), while only right wheels provide forward drive. This is to prevent accidentally sliding to the right and pushing over a ridge.

When teetering, supported only by left front, right rear. LF steer left and drive a small amount of non moving torque forward, RR steer left and drive non moving torque rearward to hold position.
RF and LR wheels steer right and attempt to maintain equal velocity independent of traction, to pull vehicle in direct steered by wheels usually off the ground. Rythmic acceleration/decelleration by RF/LR to rock vehicle so those wheels touchdown.

These "modes" wouldn't be needed by very many people, and in very few situations. But if you discover one of them helps you get through a tricky part of your "home" trail, you can set a geofence around that part of the trail so you are prompted to choose to select that mode when in that area. Other CT drivers arriving at the trailhead can receive a listing of custom drive modes generated by other drivers that have been on that trail, ranked according to popularity, and listing Driver who created the mode and the date, so you can tell your buddy, "go try out my Easter Jeep Safari custom mode on Moab Hotub." It can become a means of programming small movements in quick succession to optimize times in competition. Like video game speed runs in real life. This could work for track days too.

 

[slomobile]

The beauty of 4 independant motors with vectoring is that it doesn't matter what type of ground you are transversing. All it needs to do is match wheel speed to ground speed, whilst the acelerator controls velocity gain and steering heading change. If you want more wheel spin accelerate more like on any other car, if you want to tank turn, turn wheel and it will instantly increase outside wheel rpm and decrease inside corner rpm. They already have 3 axis sensors for ESP etc.

And if you can't steer in the direction you want to go, what other control method do you use to achieve that on a car?

I don't understand the system you are describing. "match wheel speed to ground speed" means no differential turning, no spinning wheels. Torque vectoring is not a simple calculation. It takes in to account body roll, heave, acceleration weight transfer, wheels without effective traction, and limiting torque to certain wheels in a curve to prevent exceeding traction limits and spinning out. It may feel like point and shoot when done well, but it is not simplistic. Tank turn is not just e-differential as you say. A tank turn is an abusive situation where the wheels are scrubbed aggressively. It is certainly not automatic on any vehicle with a manufacturers warranty. Outside of a Sherp maybe, or an actual tank.

Very common situation where you don't steer where you want to go is during a powerslide. Rear tires break traction in a turn causing oversteer. You counter steer to the outside of the curve and hold a fairly static angle and vary the throttle to control your direction. This is the kind of thing where ESP bumps against driver intent. An earlier post suggests other uses of torque vectoring during crawling that arent necessarily linked to steering wheel direction.

 

[JBee]

I don't understand the system you are describing. "match wheel speed to ground speed" means no differential turning, no spinning wheels. Torque vectoring is not a simple calculation. It takes in to account body roll, heave, acceleration weight transfer, wheels without effective traction, and limiting torque to certain wheels in a curve to prevent exceeding traction limits and spinning out. It may feel like point and shoot when done well, but it is not simplistic. Tank turn is not just e-differential as you say. A tank turn is an abusive situation where the wheels are scrubbed aggressively. It is certainly not automatic on any vehicle with a manufacturers warranty. Outside of a Sherp maybe, or an actual tank.

Very common situation where you don't steer where you want to go is during a powerslide. Rear tires break traction in a turn causing oversteer. You counter steer to the outside of the curve and hold a fairly static angle and vary the throttle to control your direction. This is the kind of thing where ESP bumps against driver intent. An earlier post suggests other uses of torque vectoring during crawling that arent necessarily linked to steering wheel direction.

Hey slomobile

There are various ways wheel torque and rpm can be modulated, but reality is that every single wheel experiences different forces and different levels of traction for every small section of road or track it goes over. So optimizing for each wheel individually to the particular piece of ground it is traversing at that time is necessary to achieve optimum performance.

So depending on the traction coeffcient of each wheel at that particular time, torque can be applied, and a slip ratio can be calculated to understand the traction available on each surface under each wheel. Once you have that you simply overlay the steering input to modulate the radius difference each wheel has to travel when turning, to yaw the vehicle in the direction you want, up to the physical limits the vehicle can calculate too.

Essentially, an EV with a motor per wheel, and air suspension (with load sensing, which most have) it can calculate exactly how much traction is available per wheel and command the motor controller to create the right amount of torque at the desired RPM for each wheel. On the drivers side however, the system only requires the same inputs from the driver to achieve this, being accelerate/decelerate and steer.

Because the vehicle has the capacity to understand the limits of each wheel, there is simply no need for any extra driver interaction, or other method of control because the system like that is already operating at the limit of what is physically possible for the tyres to achieve at any point on the road, and at several 10's if not 100's of calculations per second.

All of the the weight shifting from cornering, bumps, accelerating and braking, surface changes, are all boiled down to a available traction calculation per wheel that then is manipulated by driver inputs to effect the required changes commanded.

For "tank turn" you can just crank up the edifferential gain so that more rpm than calculated is given to the outside of the corner wheels etc. resulting in a powered drift, and if you want you can reverse the opposite side too. This is all very gimmicky but can use the same control method, but just amplify the control inputs.

As for your other comments on 4WS, my understanding is that the CT will have an opposite angle at slower vehicle speeds and the same angle at freeway speeds. The rear steering angle on the CT will be severely limited in comparison to the front, because the rear wheel well does not have enough space for large steering angles at all. Hence I doubt that even in a offroad situation it will have parallel capability like the HummerEV, which I generally find fairly pointless btw, and will potentially get you into more trouble than you bargin for if the user doesn't know how to use it properly, and has driven a vehicle like that for a few hours.

The reason I say this is because one of our farm tractors is a 4WS and for cropping/mowing operations we use the crab steer mode to reduce soil compaction so that the rear wheels do not follow the front wheel track. It is not as intuitive as one would think to operate it and I have never needed it offroad.

I am certain the CT will not have independent non-linked front or rear steering, so each wheel can go it's own way, simply because it will need a akerman linkage to join either side to do low scrub turns and would otherwise not be road worthy.

 

[rr6013]

Because the vehicle has the capacity to understand the limits of each wheel, there is simply no need for any extra driver interaction, or other method of control because the system like that is already operating at the limit of what is physically possible for the tyres to achieve at any point on the road, and at several 10's if not 100's of calculations per second.

As described, such a system perfectly fits snow-y and snow packed road conditions. It would introduce wheel differential speeds in a vehicular drift counter to normal, customary, and accepted steering safety practices. But such a system could win the point that human input is error if it could self-recovery from vehicle out of bounds yaw attitude.

As described, hydroplaning road conditions theoretically fit. Black Ice conditions counter the technique as described. Consistent 4 wheel minimal rotation and power maintained across time/distance contributes to foward directional stability above speed of certain minimums. Deadstop means exactly that whatever magic you bring to a Black Ice party. Pray you are not on slope.

Temperature is helpful when ambient reaches 32°F. Such a system can deactivate, returning operation to human hardware. Caveat corner case is Wyoming Interstate mountain driving condition where elevation deltas remain above 32°F BUT arctic wind flash freeze pavement with no warning. Low lying draws in Midwest operate similarly.

These exceptions define cases in which loss of traction doesn’t/shouldn’t translate to loss of power. Critically, power remaining/reduced in the absence of traction is the very essence of maintaining directional stability.

Sponsored