Blackfin:
You may be able to combine the readings from all 4, treat the chassis as a rigid body and determine a number of parameters including the angle of the uprights and the roll.
That's the sort of thing I had in mind. I don't think data from a single corner of the car could be useful on its own.
...R
How about just using existing automotive solution called ride height sensors, in other words potentiometer with a lever connected to suspension used to measure ride height for headlight or air suspension control?
When you know your dynamic camber change in relation to suspension travel (as you should in this point) you can calculate the camper based on reading of your sensors. Ofcourse one sensor/wheel is needed to count in body roll and get any meaningful results.
Or was the idea to measure/count in how much your suspension components deform? That's not possible with system described above..
MarkoY:
How about just using existing automotive solution called ride height sensors,
That looks like it might work. It is the sort of thing I had in mind without knowing that those specific sensors exist. Presumably the sensors are designed for longevity in a harsh environment.
My only doubt is whether they would be sensitive enough for the OP's project. I don't know whether ride-height adjustment needs much precision and I suspect it does not need a fast response, indeed a fast response might be unhelpful.
...R
I think there's different types of those, simplest being just kind of analog potentiometer and advanced types being digital with hall sensor and some sort of ic inside.
For example:
Continental Chassis Position Sensor
Programmable measurement ranges, e.g. 4 x 90°
Offset programmable over 360 deg
Sensitivity 0,057 V/deg nom. (programmable)
Supply voltage range: 4,75 V up to 5,25 V
Analog output, PWM, PSI5 available
Temperature range: -40°C up to 125°C
No mechanical stop
Separation of electronics and mechanics
Molded element (no PCB)
Resolution 0,022° (12 bit at 90 deg)
Designed for automotive systems (wheel housing, harsh environments)
In automotive applications those are designed to be used, the adjutment is intentionally slow (for headlight range to not go heywire when driving over a bump for example). However i think it's all done in conrol unit side. At least in analog version it's just a matter of reading it fast enough 
You see the F1 teams using those linear pot type resolvers on their suspension. Obviously they are paying thousands of bucks per corner to get the best readings.
Fortunately this technology is currently in road cars and in a better package. You don't need to waste that kind of space and mounting hardware on simple angular measurement.
I've used the MA3 encoder from USDigital. It looks like a small potentiometer but it is going to last a thousand times longer. But it is still a pain to mount it with linkages.
I believe the best solution is the Infineon TLE5012B. It is a magnetic orientation sensor. Glue a small magnet on the axis of the suspension arm which rotates (like stick it on top of a bolt head) and position the TLE chip nearby (preferably just a few mm.) It measures angles to almost 0.01 degrees and up to 2000 times per second.
One of those sensors on each corner and you're almost done. Just a lot of kinematic calculations to infer camber angle from the measured angles.
Only one thing wrong with this whole approach so far.
When a vehicle hits say a bump in the road, the suspension goes up but the vehicle stays relatively level, the whole point of "suspension".
Point 4:11 of this youtube shows what I mean.
bluejets:
Only one thing wrong with this whole approach so far.When a vehicle hits say a bump in the road, the suspension goes up but the vehicle stays relatively level, the whole point of "suspension".
Point 4:11 of this youtube shows what I mean.
Why do you think it's a problem? Any time the suspension moves there is dynamic camber change (if it's not designed to behave differently). The difference in that case is the change is happening in both sides to same direction unlike when cornering. That's why you need the input from both sides to calculate the angle of body and the tire related to track.
Maybe this helps to understand the consept: https://www.youtube.com/watch?v=aSsYWj43xBc
bluejets:
Only one thing wrong with this whole approach so far.When a vehicle hits say a bump in the road, the suspension goes up but the vehicle stays relatively level, the whole point of "suspension".
With readings from all 4 corners the software should be able to recognize when that happens.
...R
MarkoY:
Why do you think it's a problem? Any time the suspension moves there is dynamic camber change (if it's not designed to behave differently). The difference in that case is the change is happening in both sides to same direction unlike when cornering. That's why you need the input from both sides to calculate the angle of body and the tire related to track.Maybe this helps to understand the consept: https://www.youtube.com/watch?v=aSsYWj43xBc
Well aware of the operation of a McPherson strut.
In the linked video, the presenter submits his arguement based on his "simplified" diagram.
If one looks at a factory example it is sonn obvious that the geometry is completely different , essentially making the video more akin to free energy.
MorganS:
You see the F1 teams using those linear pot type resolvers on their suspension. Obviously they are paying thousands of bucks per corner to get the best readings.Fortunately this technology is currently in road cars and in a better package. You don't need to waste that kind of space and mounting hardware on simple angular measurement.
I've used the MA3 encoder from USDigital. It looks like a small potentiometer but it is going to last a thousand times longer. But it is still a pain to mount it with linkages.
I believe the best solution is the Infineon TLE5012B. It is a magnetic orientation sensor. Glue a small magnet on the axis of the suspension arm which rotates (like stick it on top of a bolt head) and position the TLE chip nearby (preferably just a few mm.) It measures angles to almost 0.01 degrees and up to 2000 times per second.
One of those sensors on each corner and you're almost done. Just a lot of kinematic calculations to infer camber angle from the measured angles.
Personally i'd use automotive sensors to automotive applications at least outside the vehicle. Surely about any solution can be made working, but if it's a matter of feasibility and durability (imagine driving in rain for example) i think it's best to use things that are actually made for exact purpose. Spending 30-100€/corner for sensors shouldn't hit race car budjet too hard https://www.autodoc.fi/topran/9706010