V.V. Vityazev

Multirate and adaptive signal processing is one of the fundamental grounds of digital information technologies existing nowadays. It can be traced back to the 70s of previous century when the formation of these grounds started independently of each other. Multirate signal processing implies the possibility to increase or decrease the sampling frequency and consequently to the change of required processing rate in the process of digital signal transformation. This fact leads to more efficient signal processing because of the possibility to decrease considerably required designed digital system computing performance. Recently much success in the sphere of multirate signal processing has been achieved. Multirate filtration and the peculiarities of its application have become the object of investigation in many scientific works devoted to digital signal processing (DSP). Dozens of studies and textbooks connected somehow with scientific and practical achievements have been published. Usage of multirate processing in the systems of adaptive and non-linear filtration, compression, analysis and decompression of speech, sound and images has unique prospects of application. Adaptive signal processing became one of the fastest developing directions in the field of DSP in late 70s - early 80s. The main feature of adaptive systems is changing in time functioning with self-regulation in accordance with given optimization criterion (adaptive filters without feedback) or training signal (feedback adaptive filters). Since 90s the efficiency of multirate processing application with the aim of adaptive filtration has been greatly developed. Mostly this fact was connected with adaptive filtration and resection of narrowband processes, but later building of signal analysis-synthesis adaptive systems using filter band banks and multirate processing became evident. A given review has the aim of making an attempt to sum up the works by the author and his disciplines in this direction since 70s of previous century till present time. The article includes three interconnected sections. The first one deals with methods of digital frequency selection of signals based on the effects of decimation in time and frequency offered by the author in the period between 70s and 80s of previous century. A fundamental idea to apply secondary discretization (decimation) and interpolation with the aim to make digital band filters and their sets was the basis to these new methods. The second section is devoted to brief description of multirate signal processing methods with the aim of adaptive filtration developed by composite author in the period between 80s and present time. The third section illustrates the efficiency of their application in radioengineering and telecommunication systems.