Uniform and non-uniform quantization of Gaussian processes
2012 (English)In: Mathematical Communications, ISSN 1331-0623, Vol. 17, no 2, 447-460 p.Article in journal (Refereed) Published
Quantization of a continuous-value signal into adiscrete form (or discretization of amplitude) is a standard taskin all analog/digital devices. We consider quantization of asignal (or random process) in a probabilistic framework. Thequantization method presented in this paper can be applied tosignal coding and storage capacity problems. In order todemonstrate a general approach, both uniform and non-uniformquantization of a Gaussian process are studied in more detail andcompared with a conventional piecewise constant approximation. Weinvestigate asymptotic properties of some accuracycharacteristics, such as a random quantization rate, in terms ofthe correlation structure of the original random process whenquantization cellwidth tends to zero. Some examples and numericalexperiments are presented.
Place, publisher, year, edition, pages
2012. Vol. 17, no 2, 447-460 p.
quantization, Gaussian process
Probability Theory and Statistics
IdentifiersURN: urn:nbn:se:umu:diva-61615OAI: oai:DiVA.org:umu-61615DiVA: diva2:570797