Digital Signal Processing Reference
In-Depth Information
Figure 12.3
Scheme of an electronic device for measuring the absolute mean value of signals
12.4
Estimation of the Absolute Mean Value
An average parameter estimation of signals can be complicated depending on
the definition of a particular parameter. It is a relatively simple matter to estimate
signal moments and related parameters, so it seems logical to begin by considering
the technique for estimating them.
A randomized scheme for estimating a signal absolute mean value is studied
here. Its performance is first of all evaluated in terms of random estimation errors.
Note that there can be instrumental as well as software implementations of this
and other estimation schemes, provided that signal quantizing is randomized or
pseudo-randomized.
12.4.1 Electronic Device
A block diagram of an electronic device for estimating the absolute mean val-
ues of pseudo-randomly quantized signals is shown in Figure 12.3. This device
comprises a pseudo-randomized ADC and some functional blocks for processing
the quantized signals. It can in fact be used to estimate two signal parameters in
parallel: the absolute mean value and the mean power. The price of adding one
more function, in this case, is very low. Only an additional simple logic gate and
adder are needed.
In accordance with the optimization results given above, the pseudo-
randomized quantizer contains only two voltage comparators, one for positive
and one for negative signal values. The quantizing thresholds
q
0
k
and
q
0
k
, pseudo-
randomly changing in time, are generated by the means of the pseudo-random
number generator and a DAC. The latter has two outputs, the first for
q
0
k
and the
second for
q
0
k
.
To ensure proper implementation of the estimation algorithm, the following
requirements should be met:
