Kalman filtering of the miniaturized inertial sensors' data for inertial navigation

The paper presents an adaptive algorithm for the statistical filtering of the miniaturized inertial sensors noise by building redundant networks of sensors in the same navigator, followed by each sensors network data fusion. The proposed method offers the advantage of having a redundant inertial navigator in terms of the detection unit. The sensors are disposed in linear redundant arrays. The novelty brought by the proposed algorithm consists in its adaptivity provided by the permanent update of the measurement noise covariance matrix [Rk] for the desired to be filtered data. In order to see how the filter works, its numerical simulation is performed by using the Matlab/Simulink software. In this way, an accelerometer sensor model is used to provide the noisy inputs. For simulation, two cases of the ideal input acceleration are considered: 1) a null signal; 2) a repeated steps signal.