Online Pulse Deinterleaving with Finite Automata

Abstract

Most of the existing deinterleaving methods work offline; they deinterleave pulse streams in a recursive way via multipass searching. Such methods do not fit online processing applications, and they have very low deinterleaving efficiency. In this paper, I address the deinterleaving problem of streams with repetitive periods, such as streams with constant or stagger pulse repetition intervals. Transitions between the pulses of different states are illustrated with regular grammars, and finite automata are established accordingly to realize online pulse deinterleaving. The states of pulse streams being deinterleaved are stored and continuously updated in the automata, and newly received pulses are judged by the associated finite state controls to determine whether they come from a certain emitter, so as to realize online deinterleaving. In the automaton-based method, multiple automata can be started to work parallel on the same stream, and they deinterleave intersected streams via one-pass (instead of recursive) searching. Simulation results also demonstrate the superiority of the proposed method in pulse-deinterleaving performances.