ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
0

How can I convert IMU sensor outputs to fuse with GPS values?

asked 2020-11-07 15:06:20 -0500

Avinash Bhat gravatar image

updated 2021-04-23 08:55:05 -0500

miura gravatar image

I am trying to develop my own localization ros package for my application. I have IMU sensor outputs and GPS sensor outputs. I want to use an extended Kalman filter to fuse and find the accurate position of the robot. But from IMU sensor, I get acceleration and angular velocity values. Whereas from GPS I get longitude, latitude, and altitude values. In order to the fusion part, both values should be of similar form. So how can I update longitude and latitude values with the help of IMU sensor values?

I am trying to develop my own localization ros package for my application. I have IMU sensor outputs and GPS sensor outputs. I want to use an extended Kalman filter to fuse and find the accurate position of the robot. But from IMU sensor, I get acceleration and angular velocity values. Whereas from GPS I get longitude, latitude, and altitude values. In order to the fusion part, both values should be of similar form. So how can I update longitude and latitude values with the help of IMU sensor values?

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted
0

answered 2020-11-09 09:55:09 -0500

Take a look at the robot_localization package. I believe it already supports your use-case. The similar form that you are looking for, and that a Kalman filter uses of course, is a Gaussian distribution. The key to the efficiency of a Kalman filter is that two Gaussians can easily be combined (convoluted).

edit flag offensive delete link more

Comments

1

I second the answer @chfritz gave. But would just note that it uses an extended kalman filter (EKF) so it can handle some nonlinearity in the models. That being said it still does treat everything like a gaussian as was stated. Cheers

JackB gravatar image JackB  ( 2020-11-09 12:56:16 -0500 )edit

Question Tools

1 follower

Stats

Asked: 2020-11-07 15:06:20 -0500

Seen: 393 times

Last updated: Nov 09 '20