Implementation of a DVB-RCS2 CPM modem simulated in a coherent environment

Abstract

DVB-RCS (Digital Video Broadcasting – Return Channel over Satellite) is an open standard for broadband satellite communication, delivering digital television and data services. In recent years a new version was released, introducing CPM(Continious Phase Modulation) to the standard. CPM has a constant envelope, i.e the transmitted power is constant. Therefor low cost non-linear amplifiers can be used, yielding excellent power efficiency.

During this thesis, the convolutionally coded CPM (CC-CPM) transmission scheme specified in DVB-RCS2[13] will be implemented in a matlab simulation environment. The transmitter is based on the Rimoldi Decomposition which separates the modulation scheme into a Continious Phase Encoder (CPE) and a Memoryless Modulator (MM).

Further a reduced complexity soft in –soft out (SISO) receiver will be implemented and performance in a coherent channel simulated for some of the waveform definitions in [13]. The demodulator performs Maximum A Posteriori (MAP) symbol detection on a reduced state trellis based on the Laurent Decomposition (LD) . The LD allows the CPM signal to be expressed fully or aproximently by Pulse Amplitude Modulated (PAM) waveforms and complex ”pseudo symbols”. For non-coherent transmissions the algorithm has to be extended to account for the unknown phase. The modelling of phase noise is a complex task and is for time constraint reasons left out of scope.

The reduced complexity receiver shows in general reasonably good performance compared to the reported performance threshold values in [14], but for the case of h=1/3 a very distinct bound on the error probability is observed. The low performance seems to be related to the state reduction in the trellis.

Further some issues on the topic of blind burst detection will be briefly discussed (frequency estimation , preamble detection etc ).