Traction Control
Contents
Rationale
Tire generates most longitudinal grip (good stuff) when slightly slipping (see pic). Need complex system of electronics and mind tricks to enable this condition at all possible times. This will enable go fast more.
This should not be confused with lateral traction which is controlled by the dark wizards in the fourth dimension who demand virginal sacrifices. Depending on the driver's skill, how much lateral load is carried by the tires, the surface conditions, the setup of the chassis etc. the targeted slip will differ.
Basic Principles
In order to ensure the the tire is working in the intended slip ratio, we first must determine what the slip of the tire is. There are a variety of ways to calculate slip but all versions basically come down to how much faster or slower the tire contact patch is traveling relative to the ground. In order to create a consistent, high performance traction control system, an accurate velocity estimation is critical.
Velocity Estimate
In a rear wheel drive vehicle, a reasonable first pass would be to take the average of the two front wheels speeds. Because these wheels are un-driven, they are generally going to be traveling the same speed as the ground. Further considerations need to be taken for an AWD vehicle sense the lack of un-driven wheels means that wheel speeds alone can't be used to determine how fast the vehicle is going. In these cases a variety of solutions such as integrating the longitudinal acceleration can be used to fill in periods where the wheels are slipping.
The same AWD considerations apply in the case of an RWD EV with regen since while braking the front wheels will be in slip.
Slip Estimation
Once the vehicle velocity is determined, it is possible to determine the amount of slip that the tire is experiencing.
It is important to also consider the effects of the different corner radii that the inside vs outside and front vs rear wheels travel while the vehicle is turning.
Slip Target
In a basic system, the slip target can be set to the slip at which the tire produces it's peak torque. If the velocity estimate is robust and the traction control regulation is well tuned, this will provide good acceleration performance in a straight line while aiding stability in a corner. Further improvements can be made by modifying the slip targets when the vehicle is turning, since with increasing slip angle the slip ratio for peak longitudinal force tends to reduce.
Combustion Engine Control
The spin of the wheels is managed by speeding them up or slowing them down. This can either be done by the engine, clutch, or brakes. Dont do brakes because thats insane [citation needed]. Clutch modulation is often present/inherent in motorcycle clutches, but do not trasnlate well to this application.
- engine
- spark
- fuel
- clutch?
- not sure if anyone does some other crazy shit like brakes or smth
EV Motor Control
EV vehicles are mostly torque controlled; i.e., the throttle pedal directly maps to how much torque the motor outputs (although there are some that are speed controlled, like boats or airplanes). Because of the fact that electric motors are much more responsive than combustion engines, traction control on electric vehicles can be quite effective. If the motor controller does not have traction control algorithms built in, a slip target algorithm can be implemented without too much headache, assuming adequate electrical engineering skills.
Wheel speed sensors must be accurate and have a high resolution. If using separate motors for each wheel, this can be done with the resolvers or encoders on the motors, but if using a mechanical differential then separate wheel speed sensors will be needed for good cornering control. Annoyingly, most wheel speed sensors are less effective at low speeds (like when launching in the acceleration event), so keep this under consideration. If you have multiple motors, or an electronically controlled differential, the traction control system will be closely linked to the torque vectoring system.
In FSAE, the traction algorithms cannot increase the requested torque, only decrease it - this might be true in FS as well. So this must be factored into your design.
In lieu of a computed traction control based on degrees of slip, there are simpler solutions. Different throttle maps can be used, to make the car more driveable. This might be nice for events like skidpad, where fine control of the throttle is more important than raw power. A simple RC filter could be inserted into the circuit, so that a fast on-throttle event results in a gradual increase to max throttle. A PID controller could also accomplish this. These solutions would need to be bypassed on off-throttle events, as the traction control algorithm cannot increase the requested torque (at least in FSAE).
Failure Modes
Since traction control is able to reduce (or completely eliminate) motor torque, it is important to consider the failure modes of the system. This can be as simple as providing the driver a switch to disable traction control, or more complex where either through electrical or model based checks of the plausibility of sensor values. In either case its possible that the driver may not realize that a loss of torque is caused by traction control, so some feedback mechanism to tell the driver when traction control is active is worth considering.
Launch Control
launch control stops burnouts which is more faster but less cool
Mid Corner Traction Control
rwd snap oversteer bad
- algo
