## Why do cars turn by turning the front wheel?

17

5

Why do we design cars to turn by turning the front wheel?

Wouldn't it better to keep all wheels straight and turn by rotating the wheels on one side of a car more then the wheels on the opposite side of the car, like tank tracks do? That way you can turn even if you are standing still. Why don't cars do this?

Karlo 2016-06-08T12:37:09.337

19

If you were going to turn left 90 degrees, without turning the wheels, then you wind up dragging the wheels sideways while you turn. 16 seconds into this video shows exactly what I'm talking about.

So every time you try to back out of your driveway, or a parking spot, or turn into a parking spot, or turn anywhere for any reason, you're going to lay down rubber because the tires are rotating quickly while remaining nearly stationary as you turn. Again, look at the wheel slip in the video I linked.

You'd be hard pressed to find "four wheel differential steering" in use anywhere for any reason. Tanks use it to reduce the ground pressure they exert so they don't sink in the dirt. Cars aren't going to sink into pavement so this isn't an issue.

1

@bcrist you could actually have tank tracks without differential steering. See such vehicles as this one: https://upload.wikimedia.org/wikipedia/commons/3/37/Tucker_sno-cat.jpg.

The primary reason is because it reduces complexity of an already complex system

Rugnir 2016-03-17T13:41:05.493

In addition to @Rugnir 's image, there are also half-track vehicles.

Chuck 2016-03-17T14:18:32.043

10The first 2 paragraphs are perfect, but everything about the last one is wrong. There are plenty of robots, forklifts, and other industrial vehicles which use 4-wheel differential steering in combination with mecanum/omniwheels to achieve 3-DOF steering (independent rotation and translation in 2 axes). Tanks do need to reduce the pressure they exert on the ground when traveling over soft terrain, but differential steering has nothing to do with it. The pressure is directly proportional to the weight of the vehicle and inversely proportional to the surface area in contact with the ground...bcrist 2015-08-30T03:59:26.797

4... thus merely having tracks/treads vastly increases the surface area contacting the ground compared to 4 wheels, decreasing the pressure imparted on the ground. But using tracks means differential steering is the only option you have. The links between segments of the tread have only a single degree of freedom; they can't "bend" to steer the tank.bcrist 2015-08-30T04:03:06.703

3My answer, including the last paragraph, is in given in reference to over-the-road vehicles. I said you'd be hard pressed to find an example, not that they don't exist. The vast majority of forklifts use rear steering, not differential steering. Mecanum wheels are too impractical to ever be used on anything other than specialty vehicles. OP's question wasn't "can anything be designed with 4 wheel differential steering", it's "why aren't cars designed with differential steering." The answer is, again, for conventional cars with conventional tires, there's too much wear.Chuck 2015-08-30T10:41:24.190

7

While it might be possible to design a rear-wheel-drive car in such a fashion that the front wheels basically operated as casters and the steering was controlled by driving the rear wheels at different speeds (some automatic guided vehicles operate on this principle), such vehicles perform poorly when there is a loss of traction at speed. If such a car were accelerating when the left rear wheel lost traction, the car would immediately start rotating to the left, and would likely travel some distance to the left before the driver could react to the situation. Even if the car's control system used accelerometers and gyros to detect that condition faster and keep the car from going far off course, they couldn't react to the loss of traction until after it had already affected the vehicle's rotational velocity. By contrast, holding the front wheels at a driver-controlled angle will keep the vehicle going straight even if one of the wheels loses traction.

If you hold the steering wheel constant, then start spinning, the vehicle won't magically go straight (although positive caster does help correct the spin mechanically). High-powered cars with one wheel on the asphalt and one wheel in the sand can be problematic for this very reason. Also, if you're steering with the rear wheels and the computer detects a spin to the left, it just has to brake the right wheel to correct the skid before the human has even recognized something happened. The problem I'd have with your hypothetical setup is the lack of redundancy. Normal cars have two places toMichaelS 2015-08-29T09:13:52.593

steer from (if the left wheel is skidding, the right wheel can be adjusted to compensate a bit), and two places to accelerate from. Even a front wheel drive car can steer and accelerate using just one front wheel, though it will be hampered. Your design means the one good wheel is "steering" by accelerating, while also trying to accelerate, which it can't do without spinning the car, so you're stuck at a constant speed unless both wheels have traction. A 4-wheel skid-steer system would be much more robust.MichaelS 2015-08-29T09:17:01.377

@MichaelS: The behavior of a non-driven passively-swinging caster is stable except when it's riding almost directly ahead of its pivot point (in which case it will sit at an unequal equilibrium until it turns one way or the other), and a such a caster can be braked when anywhere near its operating angle, though it will impart turning forces on the vehicle related to how far it is from that angle. I don't think there's a stable way for a freely-swinging caster to impart any kind of acceleration other than passive drag, however.supercat 2015-08-29T12:03:04.830

@MichaelS: Certainly it's possible to use differential drive as a supplement to angle-controlled steering wheels, but I don't think it's possible to have more than two driving wheels in a differential-drive system without either positively controlling their angle, which would be contrary to what the question was asking, or having wheels skid while turning (which is bad for reasons alluded to elsewhere. BTW, I know that a vehicle with the front wheels pointed straight ahead, one rear wheel drive, and the other not, won't go quite straight ahead (model cars with such a design need to have...supercat 2015-08-29T12:07:53.933

...the front wheel angle biased to offset the vehicle's tendency to turn slightly) but they'll go much straighter than they would if the front wheels were passive casters.supercat 2015-08-29T12:09:14.090

6

The main reason is because when a car brakes/decelerates, the load is reduced on the rear wheels and increased on the front wheels. If your car was exclusively rear wheel steering you would lose steering in high speed braking. The converse does not occur because engine accelerations are typically much much less than braking decelerations.

Forklifts are almost exclusively rear wheel steering for more maneuverability of a pallet in the front, but they are limited to low speeds.

There are four wheel steering cars: https://en.wikipedia.org/wiki/Category:Vehicles_with_four-wheel_steering

Many of them are computer controlled and steer in opposite directions for low speed maneuverability and in the same direction for high speed strafing. None that I know of do caterpillar style rotation about the center axis.

3This addresses the question in the title but not the actual question, which says nothing about rear-wheel steering. It's actually asking about steering the same way that tracked vehicles do: by having the wheels on one side rotate faster.David Richerby 2015-08-29T21:17:35.573

5

The Jeep Hurricane has four-wheel steering that allows it to pivot 360° without moving. But it's very complex and expensive.

I would guess that skid-steer systems in general are heavier, bulkier, and more complex. And really, the only place they're particularly useful is when you need really tight steering. 99% of our driving has no use for really tight steering, so it doesn't make a lot of sense to design the vehicle for the other 1%.

This line of argument is very poor. 99.99% of our driving has no use for crash airbags.March Ho 2015-08-29T09:47:57.583

2That analogy uses two things of very different impacts. On those rare occasions when seat belts and airbags and crumple zones are useful, they're extremely useful. Very few major accidents would be avoided because of four-wheel steering, which is mostly used at very low speeds to get into tight spaces better.MichaelS 2015-08-29T10:33:26.437

1

It's a versatile beast of a machine, packing lots of power, multiple attachments and (most relevantly to the topic at hand) great maneuverability.

It does exactly what you suggested. It has stationary tires, which turn at different angular velocities in order to turn the vehicle, like a tank. It's easy to turn in a full circle without changing the machine's position. This gives it a number of advantages:

• It can move a bucket sideways to smooth out dirt or gravel.
• Its turn radius is 0, meaning it can get into tight spaces not designed for vehicles.

Compared to a compact track loader (a similar machine that uses tracks instead of tires), the skid steer has these advantages and disadvantages:

• Lower maintenance and repair cost
• Lighter and faster
• Operates poorly if at all in muddy or snowy conditions
• Affects the ground more

A car has little to no need for any of these advantages, as they're not used to load, unload or landscape (though attaching a trailer would be much easier), and generally travel on pathways designed for them.

There actually exist accessories for wheeled vehicles that provide mini tracks for each wheel, to be used in mud or snow:

But even so, the basic steering functionality is still by directed turning rather than skid steering. In fact, there are even machines that are designed for tracks that steer in the same manner:

If even some tracked machines steer in the same manner that cars do, instead of by skidding, there must be some advantage, right?

One advantage (the above curbing machine steers conventionally for this reason) is that it's much easier to form a smooth curve. While computer systems could be used to determine each tire's rotation speed to produce a given turn radius, it's much simpler to simply turn the wheels. Also, as pointed out by @alphazero, one is that skidding reduces traction significantly, which can lead to uncontrollable situations at high speed. There's also the matter of wear and tear resulting from the skidding (and protip, try to turn the steering wheel in your car only when moving, if practical).

In conclusion:

Cars generally steer by turning their wheels instead of skidding because few of the advantages of skid steering have much application to a car but the disadvantages range from expensive to complex to disastrous.

1

Aside from the issue of tire wear on side-skidding wheels (which could conceptually be solved by using different materials for both wheels and road surfaces) the OP's proposed steering method has one huge disadvantage over "conventional" steering, and that is lack of precision. It necessarily involves a jump between static friction for straight-line driving and dynamic (sliding) friction when turning, and that "jump" will occur at an unpredictable point in time.

This is not much of a problem for military use of a tracked vehicle like a tank, but lack of precision steering when changing lanes when travelling at 70 mph on a busy highway is a very different matter.

0

Some food for thought. Airtrax is an omni-directional drive system with special wheels to enable the kind of maneuverability. (I've only seen it used in low speed applications)