Globally asymptotically stable filter for navigation aided by direction and depth measurements

This paper presents a navigation solution for autonomous vehicles based on direction and depth measurements, in addition to relative velocity readings, with globally asymptotically stable (GAS) error dynamics. A constant unknown drift velocity disturbance is also assumed present during the operation of the agent, which is explicitly considered in the system dynamics. The observability of the system is studied resorting to linear time-varying system theory, in an exact way, in spite of the nonlinear nature of the original nonlinear system dynamics. Realistic simulation results are presented, including measurement noise, that illustrate the performance of the achieved solution. Comparison with the Extended Kalman Filter is also carried out, revealing that similar performances are achieved for the proposed approach.