Time encoding and perfect recovery of non-bandlimited signals with an integrate-and-fire system

Time encoding represents an alternative method of sampling, based on mapping the amplitude information of a signal into a time sequence. In this paper, we investigate the problem of time encoding based on an integrate-and-fire model, consisting of an integrator and a threshold comparator. We focus on particular classes of non-bandlimited signals such as streams and bursts of Diracs, and prove we can recover these perfectly from their timing information.