This work is organized in two parts: first the statistical properties of piecewise-affine Markov maps are recalled. Analytical methods for the computation of the invariant probability densities and the rate of mixing are presented relying on the approximation of the Perron-Frobenius operator with a suitable finite-dimensional operator. Then, a modular current-mode architecture for the implementation of continuous piecewise-affine Markov maps is presented. It relies in the decomposition of continuous piecewise-affine maps by means of a suitable set of triangular basis functions which can be easily implemented by means of a daisy-chain of current-mirror based blocks. These basis functions are then weighted by R-2R equivalent ladder and summed to give the final input-output relationship. Simulations show the effectiveness of this approach in producing differently distributed and almost uncorrelated sequences.
|Original language||English (US)|
|Title of host publication||Proceedings of the 7th International Conference on Microelectronics for Neural, Fuzzy and Bio-Inspired Systems, MicroNeuro 1999|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|Number of pages||8|
|State||Published - Jan 1 1999|