Adaptive neural network filter in compensation of time interleaved A/D converter system

A time interleaved A/D converter system is a good way to significantly increase the sampling rate of an ADC for wideband shipborne software radio station. Due to errors in the manufacturing process, all ADCs in a time interleaved ADC system are not exactly identical. It means that gain and offset mismatch errors are introduced in digital signal processing. In this paper we present a method based on adaptive neural network filter for compensation of the mismatch errors. This method needs no more special information about the input signal except that it should be band limited to the Nyquist frequency for the whole ADC system. Then the compensation can be working while the ADCs are still running. Here we adopt the genetic algorithm to optimize the initial weight value in the network, this algorithm makes complementary advantages for combining back propagation and genetic algorithms. The compensation method has been validated with simulations from a time interleaved ADC system.