Good day... Sorry for bigner quesion, but really I need some friendly explanation rather than long reference I have 9 dof mpu9250 imu and gps. I found some source that calculate euler angle in ardunio How much euler angle help me in localization? Can I use the mpu9250 raw data for localization? If i fuse gps and imu what is the out put of filter? Can i determine x y z position using imu and gps?
1
-
$\begingroup$ Hi @Mjd. Welcome to this community ! Can you kindly simplify this question by trying to limit it to a single query ? It will help you get better answers faster $\endgroup$– Pe DroCommented Oct 3, 2020 at 16:32
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
I think another good place to jump right in is to examine the robot_localization package that is used in the ROS community. It implements both EKF and UKF sensor fusion estimators. The package is widely used, supported, and documented which makes it ideal for someone looking to start understanding how robot localization with multiple sensors works at a practical level. The robot_localization ROS Answers tag is very very active and is a good place to field questions about he package if you run in to them while experimenting with it.
For the particular case of implementing GPS and imu fusion look at robot_localization Integrating GPS Data.
I know this may not be a narrow and specific answer, but I think it could lead any one interested in the area onto a very useful path of discovery and has helped me build an understanding of the subject over the last year.
$\begingroup$
$\endgroup$
As an advice it's better to separate your questions and make them more precise.
In the end, if I understand correctly, your question is: how can the GPS and the IMU be used together for localization?
The IMU measures angles and angular speed but can be used as a position tracking sensor given that you know where the robot starts. You could use an extended Kalman filter to merge the sensor measurements.
I won't go in much more details because this questions has been asked before on this site and I think this answer is already an excellent read.
