Slip Angle Sensor - General Idea

[Tim.Wright]

In my experience GPS+IMU systems are pretty unreliable as slip sensors even as hideously expensive automotive specific products with sophisticated algorithms to counteract position and velocity drift coming from the IMU sensors.

The main problem is that they don't measure the slip angle directly. They are calculated based on inputs from various sensors (accelerometers, gyroscopes, GPS position + GPS velocity) which give semi redundant information and then they use a kalman filter to fuse the information together to calculate the vehicle velocity and the heading vectors then from this the slip angle is calculated.

An optical sensor instead reads the slip angle more or less directly.

[Greg Locock]

I don't know /why/ measuring the lateral velocity at each end of the car is hard without cameras, as I don't do real world instrumentation any more. But it is. i would guess that our estimates of rear slip angle using anything other than cameras can easily be out by 1 deg/g, which is the difference between really good and hmm, ok.

[Caito]

Raspberry Pi + Picam?
https://www.raspberrypi.org/blog/vector ... stimation/

I must admit "Raspberry Pi!" is my kneejerk response to just about any custom digital look-see project.

But why not? Cheap as chips and what ever you want to do, someone probably already did 80+% of it for you because there is such a high nerd-base using them.

That is indeed correct. I have been busy lately with other stuff, but will look more into detail to what would be needed to make this work.
-------------------

What about gps and a 9dof imu? The gps gives updates 5hz with interpolation from the imu.

The imu will not only give you everything you need so I can't understand why you are thinking of turning the car into a giant wireless mouse...

You could even buy a 20Hz GPS and connect it to RTK systems (to have a 1cm precision), that would give you a, let's say, precise course.

From the 9dof IMU you do some sensor fusion, smart kalman filtering, and get absolute orientation and acceleration vector. But this is very noisy, having a value that is 1deg plus/minus 2deg is not that good :p.

It will be good to tell you some noisy movements, but not good for a precise overall result. I'm talking about the less than 10usd chips you can readily buy, not complex laser cool stuff.

These are all indirect measurements, and even "double" indirects. A camera on the ground is as truth as it gets.

The idea of this "exercise" is to make something that today costs thousands, for pennies throwing some nerdism into it, to put it some way.

[Tim.Wright]

The idea of this "exercise" is to make something that today costs thousands, for pennies throwing some nerdism into it, to put it some way.

The ultimate low cost slip sensor in my opinion is this:

The threaded interface remains fixed to the body, the wheel follows the road - the angle between the two is the slip angle.

As silly as it seems I've had tests run using slip sensor which was a cross between a caster wheel and a "5th wheel":


Unfortunately I don't have images of that actual sensor but at low speeds it works better than the GPS. Kalman filters drop in reliability if the inputs are not very dynamic (low speed or extended conditions of steady state).

I have always wanted to build one of these types of slip sensors. I think for around 500€ you could have a pretty decent system. I have made a design and everything but never got around to organising it's manufacture...