Immersion and Invariance Velocity Observer Design Using Vector and Landmark Measurements

In this paper, the rotation and translation velocity estimation of a rigid object using vector and landmark measurements directly is investigated, when velocity measurements are unavailable. We consider a practical scenario that no pose information is reconstructed, then the pose estimation error cannot be directly obtained to copy with the quadratic terms existing in the pose error dynamics. To circumvent this obstacle, we provide the estimates for the vector and landmark measurements, and apply the immersion and invariance type method, which augments the state with a scaling factor. The resulted observer is proven to be globally exponentially stable via rigorous analysis. We also provide convincing simulation results to evaluate the proposed algorithm.