We combine data from different sensors like the Inertial Measurement Unit (IMU) and Global Positioning System (GPS) to estimate the state of a robot e.g. position using Kalman Filter. My question is why do we need to estimate the position if we are getting it from GPS and IMU?
This might be a matter of reliability and accuracy of the obtained sensor data, especially under real-time conditions. Additionally, it might depend on what the data is to be used for.
Hi,
It remains that the risk of failure in the IMU is not equal to zero, therefore the introduction of state estimation enhances the navigation safety of the vehicle.
For more information about this subject, I suggest you see the links on the topic.
Article A multi-sensor fusion MAV state estimation from long-range s...
Article Error Overboundings of KF-Based IMU/GNSS Integrated System A...
Article Sensor Selection and State Estimation for Unobservable and N...
https://escholarship.org/content/qt34f81319/qt34f81319.pdf?t=p7ox68
https://www.diva-portal.org/smash/get/diva2:1682155/FULLTEXT01.pdf
Best regards
Raw GPS data only provides 4 measurements of pseudo range. You need an estimator to extract the 3D position measurement and clock bias. The GPS data is then used to estimate the errors in the IMU to provide improved short-term stability for any autopilots (or a navigation capability should the GPS signals be intermittent)
- Badges
- Science topic
- Similar topics
- Mathematics