Bayesian Estimation and Tracking: A Practical Guide

20.7 Conclusions

We have demonstrated that the sensor fusion estimator produces results that are more accurate than either photogrammetry-based or IMU-based estimators alone. In the context of our UKF solver, the two types of measurements complement each other. IMU measurements mitigate the noise present in the photogrammetric data, thus making positions and angles more accurate. And photogrammetry compensates for biases and drift in the IMU measurements, making translational velocities more accurate.