regarding the existing Antilock braking system example model.

In the existing Antilock braking system example model.The wheel speed output is mentioned in unit m/s and vehicle speed output is in rad/sec .But actually when i calculated the wheel speed the output unit is rad/sec only . how did m/sec came ,there is no any block or logic near wheel speed discrete function block that converts rad/sec into m/sec . Then why there is a m/sec .The wheel speed and vehicle speed value goes into slip ratio calculation with mismatched units ,won;t this impact the accuracy of the model.please kindly provide a clarification on this.

[Fangjun Jiang]

"The wheel speed output is mentioned in unit m/s". Where is it mentioned?

[karthi]

check the wheel and vehicle speed subsystem outputs.The units are mentioned there.

[Fangjun Jiang]

What version of MATLAB is this in?

[Fangjun Jiang]

Indeed, I believe there is a mistake in the example model in R2024b.

The wheel speed signal is labeled as 'WheelAngularVelocity' and it is in the unit of 'rad/s'.

But the 'Unit' specification of the 'Signal Attributes' of the 'Outport 3' block is mistakenly specified as 'm/s'. It should be specifed as 'rad/s'.

[Fangjun Jiang]

I've submitted a service request to the Mathworks to correct this error.

[Paul]

Any idea why the incorrect unit specification doesn't cause some sort of error when building the model?

[Fangjun Jiang]

I think the unit specification is mostly for info/display only. Not all signals are consistently specified with unit. The settings for unit consistency check are probably not turned on. This is a minor error that probably nobody cares. It caused confusion for the OP though. But there are sufficient info to indicate that this is a typo error.

[Fangjun Jiang]

Mathworks tech support ackowledged the error. It has been specified as 'rad/s' in R2026a.

[Paul]

"I think the unit specification is mostly for info/display only."

The doc page (2024a) states: "By specifying, controlling, and visualizing signal units, you can ensure the consistency of calculations across the various components of your model."

I guess that wasn't the case in their example model.

I've not tried using the units functionality in Simulink but have wondered about how useful it really is.

[karthi]

MATLAB R2025b is the version I am using. Thanks for the support @Fangjun Jiang.Can you share 2026a corrected model screenshot if possible?

[Fangjun Jiang]

@Paul, "for info/display only", I meant it applied to this example model case.

Unit specification is useful, but it takes rigor modeling to reap the benefits. In the model, if I also specified the unit of the Input port of the "Relative Slip" SubSystem, then the error is shown. There are a few "Units Inconsistencies" checks available.

[Fangjun Jiang]

@karthi, to correct the error, double click the Outport 3 "WheelSpeed" block, click the "Signal Attributes" tab, change the "Unit" from 'm/s' to 'rad/s'. That is the source of the error.

[Paul]

@Fangjun Jiang

Thanks for the clarification.

Does the unit of WheelSpeed in Outport3 have to be explicitly specified as a block parameter for the Outport? Shouldn't that unit propagate to the Outport from the Wheel Speed discrete time integrator (I think that's what it is), which should automagically have the unit of its output as rad/s? Or do Simulink units always have to be set by the user and then they are only verified by the softward at subsystem boundaries (as you've shown)?

Maybe I need to explore Simulink units a bit more ...

[Fangjun Jiang]

The default setting for "Unit" is "inherit", like the "-1" for "Sample time". Yes, the best place to specify the unit is at its source. The rest of the places can be propagated. But somewhere (at the interface boundary) it needs to be manually specified, so the consistency check can be performed.

[Paul]

I looked into this and experimented a bit more.

According to Unit Consistency Checking and Propagation - MATLAB & Simulink (and what I verified), "Simulink does not propagate units through blocks that produce new data as output." For example, with a Sum block Simulink does not verify that all of the inputs have the same units (or can be automatically converted if that option is selected) and the output of the Sum block does not have units assigned. Seems to me that the primary use is to verify units at the boundaries, i.e., inports and outports of subsystems, even virtual subsystems. Not sure how it works with Model blocks.

It appears that units cannot always be specified directly at the source. For example, the Constant block does not have a block parameter to specify the units of the constant parameter. Rather, one has to use a Signal Specification block at the outport of the Constant block.

I'm still not sure about how useful this feature really is (based on very little investigation). Seems like it takes a lot of work by the user to specify the units of everything, and doing so might not be useful any way because unit consistency is not enforced (AFAICT) at the inputs to computational blocks, so user error in the specified units might not be detected.

Answer 1

[Fangjun Jiang]

Click the block masked as "?" to bring up the help document. It explains the unit of vehicle/wheel speed. There is a Gain block "1/Rr" involved to make the conversion.

[karthi]

Gain block is there "1/Rr" only in vehicle speed subsystem .But in wheel speed subsystem there is no such kind of block or conversion happening as observed and help document does not have that detail.This answer didn't clarify my doubt.Inside the wheel speed discrete integrator block parameter there is a gain,but it is not 1/Rr and the initial condition of that block is V0/Rr ,does those got anything to do with unit conversion.

[Fangjun Jiang]

From the Help document:

"The wheel rotates with an initial angular speed that corresponds to the vehicle speed before the brakes are applied. We used separate integrators to compute wheel angular speed and vehicle speed. We use two speeds to calculate slip, which is determined by Equation 1. Note that we introduce vehicle speed expressed as an angular velocity (see below)."

It is clear to me what unit is used at what place.

The input to the wheel speed discrete integrator block is "torque divided by inertia (equals angular acceleration)", which should also indicate that the unit of the wheel speed is angular.

[karthi]

You mean that both wheel and vehicle speed is in angular while going for slip calculation,right.I understood both the wheel and vehicle block is in angular ,but the unit mentioned is different. refer above screenshots of wheel and vehicle subsystem.

[Fangjun Jiang]

Correct!