Reserch On Demodulation Algorithm For Multi-h Continuous Phase Modulation

Continuous phase modulation which has constant envelop and continuous phase isa kind of high performance nonlinear digital modulation system. Because of constantenvelope, modulation signal is not sensitive to signal amplitude variation, so especiallysuitable for the high nonlinear communicarions system. At the same time it hascontinuous phase change, which reduces the signal side-lobe power, so it has the highpower efficiency and spectrum efficiency, so CPM has get more and more attention.However, its nonlinear makes the signal difficult to demodulate and difficult tosynchronize. Compared with linear modulation mode, CPM signal can improvespectrum efficiency by choosing parameters, but it also increase the complexityof the receiver. CPM receiver can through the maximum likelihood sequentialdetection docking received for the coherent demodulation signal to achieve theoptimal demodulation, and use viterbi decoding algorithm to detection. Alongwith the modulation index, modulation signal set to increase and memory lengthextension, Based on maximum likelihood sequential detection CPM ofdemodulation often needs to a large number of matched filter, existence manyphase state, The complexity of the receiver with index increased, Often, byreducing the receiver filter number, use less case shape decoder or reduce searchpath to reduce complexity, This paper is the key research how to reduce thecomplexity of the receiver.Firstly, this paper analyzed the basic principle of multi-h continuous phasemodulation, gave it signals the representation of a state and trellis, analyzedparameters on the effect of frequency spectrum and bit error performance, draw aconclusion that to get a better frequency spectrum characteristics will makedemodulation increased complexity, and gave important ARTM CPM signalparameters.Secondly, this paper studied multi-h CPM was exactly represented by a sumof pulse amplitude modulation (PAM) waveforms, and gave the simulation.Through the analysis of pseudosymbols and signal pulses, pointed out that therewas a relationship between state numbers and pulse wave length. By choosingincluding signal energy higher pulse waveform can come to approximate CPMsignal.Finally, this paper studied the optimal demodulation algorithm maximumlikelihood sequential detection basic principle, by the symstem simulation withMATLAB, analyzed of the influence of the main parameters of the bit error rate. And then based on this method, the paper proposed pulse frequency truncationoptimization algorithm. This technique can reduce the complexity of thecomplexity, and loss of limited bit error rate performance. Based on the pulseamplitude modulation (PAM) representation of multi-h continuous phasemodulation (CPM), by reducing the number of pulses and averaging the multi-hpulses, the signal was an approximate representation to reduce the use ofmatched filters, through the simulation system and compared with the optimalalgorithms, bit error rate increases a little.