Hi everyone,I'm a final year student in microelectronic engineering. I've been working on inertial measurement units and GPS for almost 6 months up to now and I think you've done a very good job there !For my part, I work with a 3 axis accelerometers, 3 axis gyroscopes and 3 axis magnetometers IMU. I use a 8 bit microcontroller from Atmel to deal with the collection of the data and it also manages the wireless communication with the base station.I've tried to implement the extended Kalman filter given in matlab files in labview using both accelerometers and magnetometers to compensate gyro's drift.I can also collect data from my gps but I wanted first to use the magnetometers to perform the correction. However, I can not get something stable and I wonder if I fill correctly my noise covariance matrix.I've calibrated my IMU and estimated the variance of the sensors estimation when the IMU remains still. In the case of my gyroscopes, I got a variance of 0.00085 along the X and Y axis and 0.003 rad²/s² along the Z(two different models of two axis gyros). For my accelerometers, I got 0.0004 g² and for my magnetometers I got something like 0.0002 gauss.From those values I should be able to fill my matrix Q, RPR and RY. However, I wonder if i do it correctly.In the case of the Q matrix, i dont understand why you used 0.00000001 for the quaternions. Could you explain that to me please ?Then , I said Q[32]=Q[40] = 0.00085 and Q[48]=0.0031. Is that correct ?Concerning RPR and RY, what I've understood so far is that I have to evaluate the influence of a variation in the acceleration over the Z estimated and the same for the X estimated. Is that correct ?I can get the yaw from my magnetometers sensing in the base frame and from my accelerometers.However, I think i'm wrong somehow cause my estimation of the euler angles from my quaternions keep on varying ...I also wonder if I could help you somehow cause I honnestly think you perform a great job and I've also a lot of project in mind (Autonomous underwater vehicle is one of the biggest)thanks anywayAll the bestYoan
You need to be a member of diydrones to add comments!
Replies
I'm to same work, I would like to know how do you make the Kalman Filter?