Position estimation using inertial sensor (dead reckoning)

Is there a filter that takes in 6DOF/9DOF IMU data and outputs position and orientation without requring GPS?

[Raj]

You will have to be more specific and clear. Sensors like AHRS can give you orientation and body velocities and you can determine the position using KALMAN filtering. GPS data will improve your accuracy in this case but it it not mandatory if you are willing to live with less accuracy.

[Swapnil Sayan Saha]

thanks!

[Swapnil Sayan Saha]

Could you provide me a matlab resource on determining pose using Kalman filter from orientation and body velocities obtained using imufilter/AHRS?

[Meg Noah]

I don't know if these will post - might be marked as spam:

Extended Kalman Filter on Git Hub

https://github.com/amelim/EKF-Slam

[Swapnil Sayan Saha]

thanks.

Answer 1

[Brian Fanous]

Hi Swapril

What your are trying to do is called “dead reckoning” and is not generally recommended with just an accelerometer and gyroscope. There are various detailed descriptions about it on the web, but NXP’s sensor fusion manual describes an experiment in section 2.8 here. If you have other non-GPS sensors that measure position, it’s possible to use the correct method on one of the insfilters to achieve your goal.