Design of an adaptive Kalman filter to eliminate measurement faults of a laser rangefinder used in the UAV system

Abstract The Kalman filter is a valued method of signals filtration having the possibility of a sensors fusion, which is commonly applied in aerospace technology. Mainly, it eliminates random noises, improves the accuracy of a measurement system and their resistance to unexpected faults. Small unmanned aerial vehicles (UAV) are a challengeable area for various applications of the Kalman filter, because their high dynamics makes them highly sensitive to external disturbances acting on the on-board sensors. The paper discusses an application idea of the Kalman filter, whose purpose is to reduce the quantity of accidental incorrect measurements reported by a miniature laser rangefinder fixed to the UAV's wing. To verify the filtration effectiveness, real distance measurements recorded during real flights were applied. The results compare two approaches: an optimization using a reference signal from the second sensor mounted to the same UAV, and an adaptation of the covariance matrix R based on innovation. We can observe that the measurements of the laser rangefinder are corrected significantly, especially for the adaptation method, what is visible as the reduced amount of the incorrect distance measurements. Hence, the reliable detection and localization of an obstacle can be achieved by the usage of the miniature laser rangefinder.