Digital Alias-free Signal Processing

12

Quantizing-specific Signal Parameter Estimation

Data acquisition does not necessarily mean the gathering of raw data. If the signals carrying information could be preprocessed in a simple and cost-effective way, then sometimes it is possible and makes sense to perform some kind of signal preprocessing carried out close to the signal sources. If the results of this preprocessing in separate data channels are acceptable for further processing of total information received from all channels then this approach is useful, as it apparently leads to data compression and less demanding requirements for the whole data gathering process. Of course, conditions for data acquisition, their processing and usage vary from case to case. Nevertheless, some typical and relatively often used signal preprocessing functions can be selected. In particular, an estimation of signal parameters seems to be a relatively popular class of signal conversions performed at the stage of signal preprocessing. They are partly discussed in this chapter and some aspects are considered in the next chapter.

A signal parameter estimation is discussed here. Imagine that an average parameter of a continuous-time signal, for instance its mean power, has to be estimated or measured by digital processing of quantized samples of the signal. To solve this seemingly simple task correctly, many fairly obvious questions have to be answered:

1. For how long should the signal be observed?
2. How precisely are the signal samples quantized ...