Harry's GPS Suite Forum

Hi Harry,

I've read that the iphone 4/4S MEMS gyroscope is pretty accurate -- have you considered generation of drift angle (yaw/slip) data? I'm not yet sure of usefulness of this data for the non-"Drifter" community although in cars such as E36, slip angles tend to be large, and might be useful. I can imagine that this data could be hard to derive and require looking at several sensors, with accuracy/update rate concerns.

Thank you,

Bill

Hi Bill,

Yes, thought about this already. I will add something like this once I integrate the Gyro in my sensor landscape.

- Harry

P.S. I'm a little afraid about the testing phase. The 911 is not nice to drift.

It's not the gyro you need for slip angle, it's the magnetometer. The slip angle will be the difference between the magnetometer heading and the GPS heading. Actually, if your GPS were precise enough, that would be all you would need. Real Time Kinematics assisted GPS (several kilobucks) is accurate to centimeters and that data is good enough to calculate slip angle as well as the travel direction. Even better if you have two RTK devices at the front and back of the car. The gyro's in a hand held device aren't really gyro's, by the way, they measure rotation rate. A true gyro gives you angles directly, you have to integrate rotation rate data to get angles and integrating the fairly high level of noise in the rate data produces a a random walk which can rapidly diverge from the true angles, certainly more than enough to accurately determine a few degrees difference between the direction the center of mass of the car is moving and the direction the car is pointing. You have to use the accelerometer and GPS heading data to continuously correct the integrated rotation rate data. But if you have a magnetometer in addition, you have additional directional information that you can use to calculate the slip angle. You don't get directional information from the GPS unless you're moving, but you're not going to have a slip angle unless you're moving.