Decoding and equalization with analog non-linear networks

Using analog, non-linear and highly parallel networks, we attempt to perfor m decoding of block and convolutional codes, equalization of certain freque ncy-selective channels, decoding of multi-level coded modulation and recons truction of coded PCM signals. This is in contrast to common practice where these tasks are performed by sequentially operating processors. Our advant age is that we operate fully on soft values for input and output, similar t o what is done in 'turbo' decoding. However, we do not have explicit iterat ions because the networks float freely in continuous time. The decoder has almost no latency in time because we are only restricted by the time consta nts from the parasitic RC values of integrated circuits. Simulation results for several simple examples are shown which, in some cases, achieve the pe rformance of a conventional MAP detector. For more complicated codes we ind icate promising solutions with more complex analog networks based on the si mple ones. Furthermore, we discuss the principles of the analog VLSI implem entation of these networks.