Can all-wheel torque vectoring replace mechanical steering?

EVfusion

Active Member
TOO Supporting Member
Joined
Mar 10, 2017
Messages
162
Location
Canberra, ACT
Country
Country
#1
The Rimac concept car has demonstrated the ability of digital torque vectoring to allow a car to mimic steering characteristics that were previously fixed because they were part of the car’s basic engineering (e.g. understeer or oversteer). The Rimac still uses hydraulic steering with turnable front wheels.
Could the Model 3 do away with mechanical steering replacing it with digital all-wheel torque vectoring? What would be the disadvantages? What would be the advantages? With so many other radical design changes included in the Model 3 it could be psychologically
a 'bridge too far'. Upfront design issues would be substantial but, if conceptually feasible, could result in further reduction in the cost of manufacturing.
Four motors would be needed to provide all-wheel drive, many additional sensors would be required and complex controls would be needed to simultaneously and individually power or brake each wheel (the Rimac controls torque to each wheel 100 times a second). Offsetting that additional complexity the suspension and power train would be further simplified and the mechanical steering components completely eliminated.
How limiting would it be to do away with mechanical steering and to steer simply by controlling the relative speed of the four wheels?
· Clearly it can be done and in simple highway driving may have some advantages. Poor wheel alignment is a major cause of tire wear. If the front wheels were fixed like the rear, alignment wear would disappear.
· Driving on poor road conditions or on unpaved roads would be much safer and significantly improved.
· But would the turning circle of the car be worsened or improved. With mechanical steering the turning circle is limited by the design chosen. As individual wheels could be stopped or reversed to facilitate tight turning (e.g. in a car park) what would determine the limits of such turns? Would there be costs in terms of increased tire wear, for example. If turning tire wear was increased, would it be more than offset by the reduction in mechanical alignment issues wear?
 

garsh

Dis Member
Moderator
TOO Supporting Member
Joined
Apr 4, 2016
Messages
10,608
Location
Pittsburgh PA
Tesla Owner
Model 3
Country
Country
#2
Could the Model 3 do away with mechanical steering replacing it with digital all-wheel torque vectoring?
Certainly not the Model 3. But, we can ask if some car in the future could do that.

As you said, you trade the complexity of a steering rack for the complexity of individual motors for each wheel, plus computer controls to interpret steering wheel input and convert it to the correct wheel speed controls for the situation. You lose the basic ability to steer when the car dies, such as to push it into a garage bay.

One awesome outcome would be the ability to pivot the car in place - right side wheels turn one way, left side wheels turn the other. The ultimate turning radius!

I'm trying to picture how parallel parking would work with such a system. I think it might have a hard time with tight turns that aren't quite a pivot. It's an interesting idea though.
 

garsh

Dis Member
Moderator
TOO Supporting Member
Joined
Apr 4, 2016
Messages
10,608
Location
Pittsburgh PA
Tesla Owner
Model 3
Country
Country
#4

EVfusion

Active Member
TOO Supporting Member
Joined
Mar 10, 2017
Messages
162
Location
Canberra, ACT
Country
Country
#5
So, this is known as "skid steering".
Sounds like the main issues are premature tire wear, and difficulty in travelling in a straight line. The later can be solved via better computer control, but the former would be unavoidable.
Pfftt...get rid of the complexity of steering altogether!
Here's a thing we made a few years ago:
Thank you @garsh and @JWardell for the feed back. "Skid steer" is what I had in mind - sorry was unaware of the term.
By having all wheels in fixed alignment mechanical steering is abolished resulting in greatly simplified mechanical arrangements. To be workable reliable and highly sophisticated individual wheel control is needed. This could be achieved in a number of ways. One approach would be for each wheel to be individually powered with precise control of torque to each wheel. Rimac have demonstrated the incredible precision that can be achieved using digital torque control to individually powered wheels. They retained mechanical steering but they could have abolished it by adding extra sophistication to their torque controller.
I agree the two main issues appear to be control (achieving reliable steering) and tire wear. But are there any other basic issues that we have missed?
  • Control: As mentioned Rimac has demonstrated that all wheel torque vectoring can deliver unprecedented precision in control of individual wheels. A standard production car could employ a simplified version of their control approach (with fewer sensors and feedback loops) but with the addition of a 'steering algorithm'. Conceptually it is very simple.
    Individually powering each wheel (non-geared in wheel motor) would be simpler in an all wheel fixed alignment car.
  • Tire wear: I'm not sure there would be premature tire wear. In fact I wonder if tire wear could be reduced. Whenever a car moves laterally relative to a straight line all wheels will suffer 'skid' wear but it is reduced for those wheels 'pointed' by mechanical steering. (With multi-wheel prime movers there are often 20 or so wheels subject to wear from this lateral friction and only 2 or 4 wheels with mechanical steering). The better road camber and speed are matched the less the wear. As a result of imperfect alignment of mechanically steered wheels, wear of the steered wheels is generally greater than for the fixed alignment (non-steered) wheels.
    I suppose the question is whether differentials and mechanical steering in conventional vehicles contribute to more or less tire wear than fixed alignment wheels and digital torque vectoring to all wheels in an alternative vehicle.
I wonder if @Mad Hungarian has a view on the tire wear issue and the trade-offs involved with an all wheel torque vectored skid steer vehicle compared to a conventionally steered vehicle. I assume these issues have been studied in depth in the context of fixed alignment multi-wheel prime movers where mechanical steering is only applied to a relatively small number of wheels.

I think the Model 3 be produced for a very considerable period and will go through a number of significant design changes. However, a fundamental redesign of the drive train would be implied by a fixed alignment all wheel drive car so I agree @garsh that if adopted it would more likely be for a future car, not a later version of the Model 3.
 

Thomas Mikl

Well-Known Member
Joined
Jun 26, 2017
Messages
214
Location
Graz, Austria
Country
Country
#6
Honestly considering the physics of steering I imagine that would be terrible for the tires and it actually may feel very unconfortable for anything than a slight bend. A 90 degree turn in a city would prolly make you cringe inside and it would be total bogus for parking.

In my book it makes no sense.
 

garsh

Dis Member
Moderator
TOO Supporting Member
Joined
Apr 4, 2016
Messages
10,608
Location
Pittsburgh PA
Tesla Owner
Model 3
Country
Country
#7
Found a paper on this. :)

http://www.ri.cmu.edu/pub_files/pub1/shamah_benjamin_1999_1/shamah_benjamin_1999_1.pdf

As the turn radius decreases from straight driving to a point turn, greater power and torque are required as a greater sideslip angle is encountered. For all turns skid steering requires greater power and torque than for explicit turning because sideslip angles are greater in all cases. In the limiting case of a point turn, the power for skid steering is approximately double that of an explicit point turn.​

So, there you go,. Skid steering is simpler, but requires more power. That also implies that skid steering is exerting a lot more force on the tires, and therefore the tires will wear out more quickly.

This paper was just for someone's Master's degree, so I wouldn't take it as the be all and end all on the subject. But it's probably legitimate, otherwise I'd expect to see more examples in the real world.
 

Thomas Mikl

Well-Known Member
Joined
Jun 26, 2017
Messages
214
Location
Graz, Austria
Country
Country
#8
The paper looks sound, however the in place turn is only theoretical as in real world conditions your rolling resistance would not suffice and you would need your wheels spinning burning rubber to do such turns. So he did simplify a lot and you would burn down tires once a week prolly.
 

garsh

Dis Member
Moderator
TOO Supporting Member
Joined
Apr 4, 2016
Messages
10,608
Location
Pittsburgh PA
Tesla Owner
Model 3
Country
Country
#9
The paper looks sound, however the in place turn is only theoretical as in real world conditions your rolling resistance would not suffice and you would need your wheels spinning burning rubber to do such turns. So he did simplify a lot and you would burn down tires once a week prolly.
He was talking about steering a robot, not a car. Quite easy to do an in-place turn when you can steer with all four wheels.

 

EVfusion

Active Member
TOO Supporting Member
Joined
Mar 10, 2017
Messages
162
Location
Canberra, ACT
Country
Country
#10
Found a paper on this. :)

http://www.ri.cmu.edu/pub_files/pub1/shamah_benjamin_1999_1/shamah_benjamin_1999_1.pdf

As the turn radius decreases from straight driving to a point turn, greater power and torque are required as a greater sideslip angle is encountered. For all turns skid steering requires greater power and torque than for explicit turning because sideslip angles are greater in all cases. In the limiting case of a point turn, the power for skid steering is approximately double that of an explicit point turn.​

So, there you go,. Skid steering is simpler, but requires more power. That also implies that skid steering is exerting a lot more force on the tires, and therefore the tires will wear out more quickly.

This paper was just for someone's Master's degree, so I wouldn't take it as the be all and end all on the subject. But it's probably legitimate, otherwise I'd expect to see more examples in the real world.
Thank you @garsh - an interesting paper.
Yes, I think we can safely conclude that it is not practical to apply all-wheel torque vectoring to four fixed alignment wheels as an alternative to mechanical steering. The lateral forces resulting from skid steer would require heavier suspension, result in additional tyre wear and require additional power consumption. In an automotive context (in possible contrast to specific robotic transport requirements in an interplanetary context) the gains from fewer parts and simpler construction would be more than offset by the identified losses.
 

Mad Hungarian

Resident M3OC Wheel/Tire Guru
Joined
May 20, 2016
Messages
909
Location
Montreal, QC
Tesla Owner
Model 3
Country
Country
#11
Thank you @garsh - an interesting paper.
Yes, I think we can safely conclude that it is not practical to apply all-wheel torque vectoring to four fixed alignment wheels as an alternative to mechanical steering. The lateral forces resulting from skid steer would require heavier suspension, result in additional tyre wear and require additional power consumption. In an automotive context (in possible contrast to specific robotic transport requirements in an interplanetary context) the gains from fewer parts and simpler construction would be more than offset by the identified losses.
Sorry to chime in so late here, have been burning the midnight oïl both at work and home lately...
In any case it looks like the crew here have already beaten it to death, as it should have been. Skid steering is clearly not a good idea for a long range road EV for the reasons stated, however torque vectoring is hugely useful in improving handling as you can introduce all kinds of corrective forces that chassis tuning alone can't do, and you can alter it virtually in real time based on conditions and driver preferences. I'm a huge fan. What I'm NOT a fan of is that Tesla won't let us turn off the ESC (electronic stability control) and fully enjoy the remarkable balance and agilities of their chassis. Virtually every other high performance brand allows this. But that's a rant for another thread.