Automatic variance control and variance estimation loops

Abstract

Abstract.A closed-loop servo approach is applied to the problem of controlfing and estimating variance in nonstationary signals. The new circuit closely resembles, but is not the same as, automatic gain control, which is common in radio and other circuits. The closed-loop nature of the solution to this problem makes this approach highly accurate, and it can be used recursively in real time.

Control systems approach to the sample inverse covariance matrix

“…1 is a multivariable version of an automatic variance control loop discussed elsewhere [7]. The identity matrix shown has dimension equal to the number of unknown parameters (or weights), taken to be m. The set-point scalar r ¼ 1 and K (not to be confused with the lowercase subscript which is the discrete time index) is a matrix of gains, normally taken to be a diagonal matrix for simplicity.…”

Single temperature sensor superheat control using a novel maximum slope-seeking method

“…A variance based control method was investigated in Vinther et al (2012a), where it was discovered that the variance of the outlet temperature increased at low superheat, which can be used for feedback purposes. This is also closely related to the automatic variance control method (Moir, 2001), which in some cases has been used in the conventional two sensor superheat control to adjust the reference. A problem with the variance based method is its sensitivity to operating conditions and to large disturbances, as this can change the variance level.…”

Filtering, smoothing and prediction using a control‐loop spectral factorization method for coloured noise

“…APPENDIX A: SIMPLE VOICE ACTIVITY DETECTOR An estimator for inverse variance has already been discussed elsewhere [18]. The schematic is shown in the diagram next.…”

“…The estimators (16)- (18) in the preceding section are of little computational use as they require the removal of the [. ]+ brackets.…”

“…The factory noise was obtained from the Noisex corpus and is signal 021. A simple VAD was employed based on smoothed variance [18,19] (see Appendix). The SNR over the whole sentence was measured as 8.6 dB.…”