Self-routing analog-digital converter based on two-layer neural network

In this article structure of self-routing analog-digital converter based on neural network is described, this structure contains three blocks: block of comparators, ADC length testing scheme and neural network. Input signals proceed into block of comparators and ADC length testing scheme. ADC with adjusted length for any input signals depending on necessary number of ranks. Also two-layer neural network structure is described. It includes input and output layers, switches and basic neurons. Switches provide self-routing algorithms into neural network and basic neurons provide analog-digital conversation. Basic neurons are one-rank ADCs. Work of ADC is three stages: creation of individual ADC for input signal, analog-digital conversation and destruction of individual ADC. Individual ADC creation is association of basic neurons. Additional ways between neurons increase fault-tolerance, because they let exclude disabled neurons. Analog-digital conversation implements by bit-by-bit method, every rank works on two times.