Dynamic Land-based LiDAR Systems: IMU Integration
Researchers: Youssef Kaddoura, Ahmed Mohamed
Description: Develop a technique that will integrate IMU and Lidar using Kalman Filter estimation approach. The catch to IMU-Lidar are the errors from sensors. Although the error can be small - for some accelerometers- it is nontrivial. Especially when you integrate for a long period of time, the noise starts to accumulate and it show up slowly at the beginning but then very fast drift, slowly at first, but they're cumulative. The contribution will be building a model that help minimize the errors, by exploiting the Lidar capability. Usually, Lidar give out laser data that can contain some known targets, and feed them into Lidar to null the attitude errors. So by scanning known targets can give a good idea of which way your vehicle is pointed. Then these corrections will be input into Kalman Filter as an update. By feeding the trajectory and the known points positions to the model, then the system will know when to identify targets and do the drift corrections.
Documents: