Kalman filter estimation with edge detection-based hybrid sensing
Yige Chen, Kenn Richard Oldham · 2016
A novel sensing and estimation scheme is introduced as an integrated part of a micro-stage based calibration method for a MEMS gyroscope. The proposed sensing scheme combines a periodically updating analog sensor with an event-based threshold sensor. A first derivative of Gaussian filter is applied to the threshold sensor signal to detect threshold crossings. A Kalman filter with the proposed sensing scheme is used to estimate the micro-stage's states. In the measurement update steps, analog sensor measurement is used when a threshold crossings event is not triggered, and the threshold sensor measurement is used when the crossing event is triggered. The impact of detection error of the proposed threshold detection technique on the estimation performance is evaluated with estimation error covariance analysis and simulations.