Motorcycle Traction Control: Grip, Slip and the Torque Between Them

Some wheelspin is not necessarily a problem

A driven tyre can generate useful longitudinal force while operating with some slip.

So the objective of performance traction control is not necessarily:

Zero wheelspin.

A simple representation might be:

Slip % ~= (rear-wheel speed - reference vehicle speed) / reference vehicle speed x 100

If the motorcycle is travelling at an estimated 100 km/h while the rear tyre's circumferential speed corresponds to 110 km/h, we've got roughly 10% positive slip.

Easy.

Except motorcycles have a habit of making easy things awkward.

Effective tyre radius changes with load, temperature, speed and lean angle.

The front wheel unloads under acceleration.

Eventually it may leave the ground entirely.

At that point, treating front-wheel speed as an unquestionable measure of vehicle speed becomes optimistic.

Modern traction-control systems therefore need considerably more context than:

Rear wheel faster than front wheel = panic.

Wheel speeds can be combined with IMU information, lean angle, acceleration and other estimates of motorcycle state.

The useful question is not simply:

How much slip have we got?

It is:

How far are we from the slip we actually want?

First, make sure the wheel speeds tell the truth

Before asking how clever the controller is, ask whether its inputs are right.

Wheel-speed scaling depends on tyre dimensions, sensor configuration and the reference-speed model being used.

If that setup is wrong, the ECU can see a slip error that comes from the model rather than the tyre.

Filtering matters too.

Filtering can remove noise and spikes from a wheel-speed signal.

Too much filtering can also hide the beginning of a real traction event.

Which is an awkward way to start a conversation about reacting early.

Think of slip like a wildfire

This is the analogy I tend to use with riders.

Slip is like a wildfire.

Catch an ember early and it takes very little effort to stamp it out.

Leave it alone until half the forest is burning and suddenly you need a considerably bigger intervention to regain control.

A traction-control event is much the same.

If the controller reacts early to a developing slip error, only a relatively small torque correction may be required.

Let the rear wheel accelerate significantly beyond the target and the controller has a much bigger problem.

More torque has to be removed, often more quickly, and the intervention becomes far more obvious to both the rider and the tyre.

That leads directly into my preferred calibration philosophy:

Active early. Intervene lightly.

I'd rather have a controller making small corrections early enough that the rider barely notices than one that waits for a large slide and then arrives carrying a sledgehammer.

Good traction control does not need to sound impressive.

It needs to work.

What does traction control actually control?

Ultimately:

Engine torque.

Grip is the constraint.

Torque is what we can manipulate.

The rider asks for something through the throttle grip, but on a modern ride-by-wire motorcycle that request can pass through several layers before it becomes rear-wheel torque.

Grip position does not necessarily equal throttle-butterfly position.

Throttle-butterfly position does not necessarily equal engine torque.

And requested engine torque does not necessarily equal the torque the ECU finally allows.

That gives the controller several possible ways to intervene.

Electronic throttle can reduce airflow.

Ignition timing can be retarded.

Individual combustion events can be cut.

Fuel can be cut.

Different systems use different combinations.

The engineering problem is not simply:

How do we reduce power?

It is:

How much torque do we need to remove, how quickly do we need to remove it, which actuator should do the job, and how do we put the torque back afterwards?

That last part matters.

Remove too little and slip continues to build.

Remove too much and you kill drive.

Return the torque too aggressively and you can immediately create another traction event.

Return it too slowly and the rider spends the rest of the corner waiting for the motorcycle to remember it has an engine.

The target tells us what we want. The strategy decides how we get there.

A slip target is only one part of the system.

It defines the behaviour the controller is trying to maintain.

The controller still has to decide what to do when measured behaviour moves away from that target.

How large is the error?

How quickly is it growing?

What gear are we in?

How much torque is the rider asking for?

How much lean angle is present?

Is the motorcycle still at large lean or already being picked up?

Has the controller already intervened?

Did that intervention work?

This is where terms such as proportional gain, integral action, derivative action, PID tuning, gain scheduling, torque-limit tables, filters, multipliers and actuator strategies start appearing.

We're deliberately not going much further into those here.

Because this is the point where:

"How does traction control work?"

turns into:

"How do I actually calibrate one?"

That is a different article.

And probably a course.

Not all traction control is the same

A lot of discussion around motorcycle electronics gets muddled because very different systems are all called traction control.

Selecting TC level 3 on a road bike is not the same job as configuring BMW M Race Calibration.

And neither is the same as working inside a full competition ECU where the controller behaviour itself is exposed to the engineer.

It helps to think of three broad levels.

Level 1: Production traction control - the safety net

Production motorcycle traction control can be extremely sophisticated internally.

But its design brief is broad.

It has to work for different riders, tyres, temperatures, surfaces and conditions, usually without requiring specialist knowledge from the person riding the bike.

From a racing point of view, I largely think of this as the safety-net layer.

Its job is to catch a situation heading somewhere undesirable.

Save the bike.

Save the skin.

Give the rider another chance.

That's valuable.

It just isn't the part we're particularly interested in as race engineers.

Typically, the rider gets an intervention level or a riding mode.

There may also be independent adjustments for things such as wheelie control, engine braking and throttle response.

But you're normally selecting the behaviour provided by the manufacturer.

You're not engineering the control strategy.

For racing, we're interested in moving beyond the safety net and using electronics as a performance tool.

Level 2: OEM race-kit electronics - useful authority without having to invent the controller

Move into manufacturer race-kit electronics and the design brief changes.

BMW M Race Calibration, Honda HRC kit electronics and Yamaha GYTR/YEC are good examples of this middle ground.

These packages are designed for race engineers, crew chiefs and data engineers who need meaningful control over motorcycle behaviour without having to develop the control system from scratch.

BMW's M RCK is a particularly useful example.

Here, traction control is no longer just a rider-facing "more or less TC" setting.

The system operates as a closed-loop slip controller.

Measured rear slip is compared with a state-dependent target, and the system requests torque reduction when the measured behaviour departs from what is wanted.

The engineer can influence target behaviour, rider offsets, intervention character and how that control changes through the lap.

That is genuine race calibration.

But BMW has already done the architecture.

The controller exists.

The vehicle-state logic exists.

The available actuators have been selected.

The calibration boundaries have been defined.

The engineer is tuning the manufacturer's system rather than designing traction control from first principles.

For most race teams, that is exactly what you want.

It lets you concentrate on making the motorcycle faster rather than spending Friday practice discovering that Thursday night's controller change has turned corner exit into a university research project.

Honda HRC and Yamaha GYTR/YEC packages follow the same broad principle.

More meaningful race calibration than a production ECU.

Far less freedom - and risk - than a completely open competition strategy.

For a lot of professional racing, that's the sweet spot.

Level 3: Full competition electronics - now you need to know what you're doing

Then we reach the more open motorsport platforms.

Cosworth.

Bosch Motorsport.

MoTeC.

Marelli.

Mectronik competition applications.

Here the question can move beyond:

What slip target do we want?

towards:

How do we want the controller itself to respond?

Depending on the platform, the engineer may gain access to things such as:

P, I and D gains.

Gain scheduling.

Controller multipliers.

Slip estimation.

Filtering.

Torque-limit structures.

Ignition intervention.

Throttle intervention.

Cut strategy.

Torque recovery.

State-dependent control logic.

At this point you aren't simply moving a traction-control slider.

You are doing control-systems engineering on a racing motorcycle.

You don't literally need an engineering degree before the software opens.

Understanding control theory before changing everything in it is still probably advisable.

More calibratables do not automatically mean better traction control.

They mean more authority.

And authority works both ways.

A skilled engineer can produce an intervention that happens early, removes only the torque required and is barely noticed by the rider.

A poor calibration can turn a perfectly good motorcycle into something that cuts cylinders, shuts throttles and argues with the rider all the way out of the corner.

Same motorcycle.

Very different outcome.

Bosch and MoTeC: when the internal structure becomes visible

Published motorsport strategies give us a useful look at what is happening beneath the rider-facing TC number.

Bosch Motorsport systems expose concepts such as slip set-points, controller gains, correction maps and intervention limits.

MoTeC motorcycle strategies similarly expose slip-based control with configurable controller behaviour and state-dependent compensation.

The vocabulary changes.

You're no longer just discussing "more TC" and "less TC".

You're discussing:

Target. Error. Gain. Response. Torque reduction.

It is the same control loop we started with.

The engineer simply has much greater access to it.

Marelli: one motorcycle, several controllers

Marelli's long history in top-level motorcycle racing also illustrates another important point.

Traction control does not exist in isolation.

A modern racebike may have several functions influencing engine torque at the same time.

Traction control wants a reduction because rear slip is too high.

Wheelie control may want a reduction because pitch or front lift has exceeded its target.

Engine-brake control is managing negative rear-wheel torque on entry.

The rider is asking for something else again.

The clever part isn't simply having each function.

It is making the whole control architecture work together without the rider feeling like five engineers are fighting over the throttle cable.

Mectronik

Mectronik also belongs in the modern race-electronics discussion, although detailed public information on its higher-level motorcycle traction-control strategy is comparatively limited.

For that reason, there is little value in inventing calibratables that haven't been publicly documented.

What matters here is the wider architecture.

Modern competition systems increasingly separate the desired tyre behaviour from the controller response and from the engine actuators used to achieve it.

As more detailed Mectronik strategy information becomes publicly available, it will be interesting to see exactly how much control is exposed to the customer engineer and how much remains within a predefined race-kit framework.

Cosworth: the control system should match the motorcycle

Cosworth gives us a particularly good historical example of why traction-control sophistication should match the performance problem.

Ten Kate Honda used the Pectel/Cosworth SQ6 architecture on its CBR600 World Supersport machinery.

The higher-powered World Superbike application used the more capable MQ12 architecture.

That distinction makes sense.

A roughly 150 bhp Supersport motorcycle does not present exactly the same rear-tyre control problem as a full Superbike.

There is less outright engine torque to manage and less need to add complexity simply because the ECU is capable of providing it.

But the underlying control problem remains the same.

Define the behaviour you want.

Measure what is actually happening.

Compare the two.

Decide how strongly to respond.

Reduce torque.

Then return the torque properly when the event is under control.

The more capable Cosworth strategies expose increasingly detailed control over that process.

Slip target and controller response are separate things.

So are intervention and recovery.

That distinction matters.

The target says what behaviour we want. The gain determines how strongly we respond when we move away from it.

And that takes us straight back to the wildfire.

Two traction events can ultimately reach the same amount of wheel slip but develop in completely different ways.

One may gently move beyond the target.

Another may have the rear wheel accelerating away rapidly.

The final amount of slip might look similar.

The control problem isn't.

A sophisticated controller can respond not just to how large the error has become, but to how quickly that error is developing.

That is where proportional, integral and derivative behaviour - and rate-based terms more generally - become useful.

We don't need to tune them here.

The important idea is simply this:

Rate matters.

Once the controller decides torque must be reduced, there is still another engineering decision beneath it.

How should that torque reduction happen?

Ignition retard?

Cylinder cut?

Throttle intervention?

Some combination?

And once the tyre comes back under control:

How quickly should we give the torque back?

Catch the slip.

Control it.

Return the drive properly.

That is the difference between having traction control and having well-calibrated traction control.

Does traction control automatically make a professional rider faster?

Not necessarily.

A top-level rider already has an extraordinarily capable feedback-control system.

Eyes.

Hands.

Inner ear.

Seat.

Feet.

Experience.

And a very well-developed right wrist.

Elite riders can feel rear-tyre behaviour and modulate engine torque to a remarkable level themselves.

Simply fitting traction control therefore does not guarantee a faster single lap.

Where a well-calibrated system becomes particularly valuable is repeatability.

The tyre changes.

Fuel load changes.

Grip changes.

Track temperature changes.

The rider gets tired.

A controller can reduce how much of the available performance depends on the rider manually managing torque perfectly through every one of those changes.

That can reduce workload and improve consistency.

It can also narrow some of the advantage enjoyed by riders with exceptional throttle feel.

It does not replace rider skill.

The rider still has to load the tyre.

Pick the line.

Finish the corner.

Get the motorcycle pointed.

And know when to ask for drive.

Traction control cannot rescue a rubbish corner exit.

We're engineers, not magicians.

It can also change the engine you can afford to use

This is where TC becomes particularly interesting from an engine-development perspective.

Imagine an engine with an aggressive torque delivery.

Perhaps it makes excellent top-end power but is difficult to use cleanly at lower speed or larger lean angle.

One solution is to soften the engine everywhere.

Reduce torque.

Flatten the throttle response.

Make the whole thing gentler.

Sometimes that is absolutely the correct answer.

But it can also mean compromising the engine everywhere because it is difficult to use somewhere.

A properly developed torque-control strategy gives us another option.

Keep the aggressive engine where the tyre can accept it.

Manage the delivered torque where it cannot.

Traction control still hasn't created any grip.

If the tyre cannot transmit the requested force, the ECU cannot manufacture the missing friction.

What it can do is reduce torque demand until the tyre is being asked for something it can actually deliver.

Or, put less formally:

A 220 bhp engine is only impressive until the rider spends half the lap manually converting it back into 160.

Good torque control lets us be cleverer about where that performance is used.

Good traction control shouldn't need to shout

This brings us back to the wildfire.

The most impressive traction-control system is not necessarily the one you can hear cutting cylinders from the other side of the circuit.

Heavy intervention may simply mean the system arrived late.

Catch the developing slip early and the correction can be small.

Let the error grow and the controller needs progressively more authority to recover it.

That is why I generally favour an active system with relatively low intervention gain.

Small correction.

Early.

Stable motorcycle.

Keep the tyre near where you want it.

The rider may barely notice anything happened.

And that does not mean the traction control is doing less.

It may mean it is doing its job properly.

The important distinction

Traction control does not create grip.

It cannot increase the coefficient of friction.

It cannot produce tyre load that isn't there.

And it cannot negotiate with physics.

What it can do is manage the torque demand placed on the tyre.

The more sophisticated the system becomes, the more intelligently it can answer four questions:

What is the tyre doing?

What do we want it to be doing?

How much torque should we remove?

How should we remove it?

That is the foundation.

Everything after that is calibration.