A time-variant noise estimation algorithm for channel observation in sensor systems
Herbert Nachtnebel, Thilo Sauter · 2006
If a measurement principle allows the active stimulation of the signal to be measured in a quasi-closed-loop setup, a pass-band measurement approach is preferable to base-band measurement since it allows changing the carrier frequency to avoid disturbances in the case of electromagnetic interference. To assess the signal-to-noise ratio on the measurement channel, a noise estimator can be used to gather the channel statistics. Although a straightforward implementation of a Gaussian variance estimator yields a feasible implementation, we show that improvements of the estimator can be made which both reduce the implementation complexity and improve the system behavior under normal, low noise, operating conditions without sacrificing the time needed to react to disturbances.