How does AP differentiate human torque from wind gust?

[insaneoctane]

Really windy in Southern California this time of year. I have been really appreciative of how well AP maintains lane position during huge gusts of wind. Possibly better than me. Makes dicey conditions far more comfortable....

But thinking about this made me wonder how AP knows the difference between me tugging the wheel to take control over AP vs a huge wind gust? If I wasn't using AP, a huge wind gust would certainly jerk the wheel, no? It doesn't take a huge amount of torque to take control. Obviously if a gust of wind accidentally canceled AP when my hands were not on the wheel that could be very bad!

 

[discoM3]

If your hands were not on the wheel, then no torque would be applied to the wheel. So nothing sensed.

 

[earthwormjim]

Unless your windows are down and you strapped a sail to your steering wheel, a gust of wind does not apply any significant amount torque to the wheel.
Wind doesn't specifically apply a torque on your front wheels anyway, it's blowing the whole car across the road. Being dragged across the road is what applies a small amount of torque.

A very very small amount of that torque would be transmitted by the redundant mechanical linkage in the steering column, but it's not enough to even be felt. It might not even show up at the steering wheel at all, instead it might just flex the steering coupler a little.

What you are actually feeling is merely the drive by wire system transmitting feedback to you, it's not "real" torque from the wheels, it's all manufactured. The car is sending you that sensation of the wheel moving, not the other way around. Since it's the car that is triggering the feeling of torque on the steering wheel, it of course knows its not your input triggering it.

Going further, what about a major pothole or obstacle which does cause a significant amount of wheel turning? I wouldn't be surprised if the car could account for that. The steering column has a flexible coupler, so sensors could easily detect which side of the coupler triggered the movement. Did the rack move first, or the steering wheel? All drive by wire systems need to be able to account for this, otherwise you will have positive feedback, road causes the tires to turn slightly, tires apply torque to the steering wheel, drive by wire systems pick up this torque, and boost it, applying more torque to the wheels. Can't allow that to happen.

If your hands were not on the wheel, then no torque would be applied to the wheel. So nothing sensed.

There's still a mechanical link between the steering rack and the steering wheel. If torque is applied to the wheels, a tiny amount gets transmitted to the steering wheel.

 

[diplomat33]

But thinking about this made me wonder how AP knows the difference between me tugging the wheel to take control over AP vs a huge wind gust? If I wasn't using AP, a huge wind gust would certainly jerk the wheel, no? It doesn't take a huge amount of torque to take control. Obviously if a gust of wind accidentally canceled AP when my hands were not on the wheel that could be very bad!

Wind only applies a bit of force to your car causing the car to be pushed out of the lane a little, not the steering wheel. Wind does not apply any significant torque to the steering wheel.

 

[MentalNomad]

[bad answer removed]

 

[Occar]

The steering is "drive by wire." Since there's no steering column mechanically connected to a steering rack, a wind gust will not make the steering wheel jerk - it would possibly lean the car and make the tires turn, but that won't affect the wheel (unless the car turns the steering wheel to compensate.)

This is most certainly not the case. It's normal electric power steering, but programmed to do more than just assist as it will steer itself. If I recall correctly, the motor isn't any beefier than in normal electric power steering either. I know on the Model S it is (or was) the same motor used in EPS systems in other cars.

 

[MentalNomad]

This is most certainly not the case. It's normal electric power steering, but programmed to do more than just assist as it will steer itself. If I recall correctly, the motor isn't any beefier than in normal electric power steering either. I know on the Model S it is (or was) the same motor used in EPS systems in other cars.

I stand corrected. If it's not to late, I'll edit my post.

 

[Occar]

To answer the question asked by the poster, there is a torque sensor on the steering wheel side of the assist motor. This would allow the car to know if torque is applied by the steering wheel, or forces outside the car. For example, if there is a torquing force being applied to the motor, but the torque sensor on the steering wheel side measures no torque, then there isn't any steering wheel input, and the motor should resist that force (it's the same as power steering setups that have crosswind or road crown compensation. If the torque sensor detects torque, and the motor also detects torque, then it should assist (or disengage in the case of AP).