Lateral acceleration on a motorcycle with RaceBox Mini fix

privilege15

Hello.

I've done a bit of testing with the latest RaceChrono Pro version (8.0.8) on a motorcycle where the data is collected from the IMU of the RaceBox and everything looks accurate except for the lateral acceleration figures.

The matter is, while longitude acceleration is quite accurate when it's obtained from the RaceBox itself, the lateral acceleration from the RaceBox cannot reflect the true value due to the specifics of how motorcycles turn. In the previous versions of RaceChrono Pro where lateral acceleration was calculated based on GPS data, it was much closer to be accurate.

While I am not an expert, I can tell that my lateral acceleration never exceeded 0.2G in the new RaceChrono Pro version no matter how fast I negotiated turns and how much I leaned the motorcycle. That includes leaning my motorcycle up to 44 degrees, it was still showing no higher than 0.2G which can't be true. In the older version with data calculated based on GPS input, it used to reach 0.6-0.7 on a regular basis and sometimes even higher, while riding in the twisties.

While searching the internet on motorcycle data acquisition, I found this quote at https://www.datamc.org/data-acquisition/g-forces-and-acceleration/ , which sheds some light on potential reason for the false reading on the motorcycles: "...In this section we will discuss both longitudinal acceleration (which includes braking) and lateral (side-to-side) acceleration.

The use of a GPS-based system is imperative if lateral acceleration is to be considered. A non-GPS-based system is incapable of providing lateral acceleration data without some very elaborate and expensive equipment."

Now here is my question, in the vehicle profile section, when we choose "motorcycle", could you change lateral acceleration calculation method from RaceBox Mini IMU back to GPS based or at least give us an option to choose lateral acceleration from being taken from RaceBox accelerometer back to GPS based calculated figure.

aol

The v8.0.8 introduced IMU lean angle functionality which requires the RaceBox Mini to attached on the motorcycle for the duration of the session. You can go to RaceChrono > Sessions > (your session) > (top menu) Session settings > and switch the IMU source to "GPS", to restore the previous theoretical GPS functionality.

privilege15

IMU is useful for the accuracy of the lean angle so it would be counter productive to turn it off entirely.

Here is the the comparison screenshots taken a few months apart on the same road trip: https://ibb.co/ySsZHcK. Notice how lateral acceleration readings changed and in the latter the lateral acceleration is basically non existent.

The RaceBox is attached to the bike as always. The only major difference between these sessions is the RaceChrono version. It is great that the lean angle is now read from the IMU in the latest version, but we should be given an option to read lateral acceleration from the GPS. If we switch to GPS in the session settings, it switches everything to GPS and we loose the IMU precision of the lean angle then. I would assume it's everyone's issue now who rides motorcycles.

privilege15

Here is what I mean. If I switch the session settings from IMU to GPS, take a look at these two screenshots: https://ibb.co/qkCxnFm

The IMU screenshot reflects accurate lean angle but false lateral acceleration.

The GPS screenshot reflects accurate lateral acceleration but false lean angle.

Both screenshots are taken from the same session file and same location but with different settings.

You see, motorcycles, unlike cars, have to lean to make a turn, thus RaceBox is never horizontal for the duration of a turn, which I would assume, turns out in a false lateral IMU reading because the lateral force is pushing the RaceBox down instead of sideways as in a car so it "thinks/senses" as if nothing is happening. In a car, the RaceBox (or any other sensor) is always horizontal, so the lateral force pushes it from the side and the IMU reading is always correct there unlike on a motorcycle.

So the problem that I can see is the reading of a lateral acceleration only, which cannot be interpreted equally between two- and four- wheeled vehicles simply due to the nature of how both corner. The longitudal reading, on the other end, is fine, because the position of the RaceBox against the horizon doesn't change during acceleration and braking in either type of vehicles.

privilege15

We are not asking much in terms of application changes. Just giving us the option in the session settings to change lean angle and acceleration sources separately, such as:

1. Lean angle source: IMU/GPS

2. Acceleration source: IMU/GPS

And that's it. This will make it very flexible and everyone can choose the source separately to his/her liking. For example, lean angle - IMU, acceleration - GPS or the other way around or both GPS, etc...

Quentin

Did you contact Racebox about the issue with wrong lateral accelaration, they advertised it as a feature so should work on it

privilege15

Quentin Did you contact Racebox about the issue with wrong lateral accelaration, they advertised it as a feature so should work on it

No I did not contact them, because I'm not so sure it's the fault on RaceBox behalf or an issue. As far as I have read and understand the topic, it's just impossible to implement via IMU, which is attached to the bike, uness it is attached to a gimbal first, which would keep it horizontal all the time but that's not practically possible and it wouldn't be showing lean angles then :) So the GPS based calculated acceleration readings is the only way to get it right while on a motorbike.

Again I'm no expert, I'm just using some common sense, logic, data from the Internet and actual data collected from my bike twisties sessions. If I'm not right, someone please correct me.

So if we could have the option to use IMU for lean angles and GPS for acceleration figures, that would be already pretty neat.

aol

@privilege15 I think you're expecting calculated lateral acceleration and real lean angle. You may call it false, but when you think about it, RaceChrono does it correctly. The lateral acceleration is in relation to the bike, so basically it's almost always near 0 G.

CTK

I know this is an ancient thread but I was googling for a solution and just figured one out

aol's post is correct and good insight. IMO it is better to just use calculated GPS data. From what I've seen Racebox's on board lean angle calcs are kind of optimistic. The only places I've seen GPS data get a little weird is if a corner has banking or big elevation changes. Since the GPS "lean" is back calculating from your acceleration (Gs = tan(LA) and vice versa) if a corner gives you grip that will translate into a higher than reality lean angle. It will never be perfect but it's good to understand what data is coming from where.