Sequential sequence estimation for CPM

Gerhard Kramer, Ed Shwedyk

Research output: Contribution to conferencePaperpeer-review

Abstract

A major disadvantage of CPM is the complexity of the optimal receiver. The number of states in the trellis representation are on the order of M, where M is the modulation alphabet size and L is the memory imposed by the modulation scheme. In this paper we investigate the reduction of the graph search by using a sequential search. Simulation results show that the bit error probability for the M-Algorithm is considerably larger than its error event probability due to the tendency for long error events to occur, an effect found by Simmons to occur generally for h=k/p when k>1. Due to its early decision mechanism the M-Algorithm tends to false lock onto paths which are 2π±Δφ away in phase; Δφ is a small phase offset. The delayed decision mechanism under which the SA operates neutralizes this phenomenon.

Original languageEnglish
StatePublished - 1994
Externally publishedYes
EventProceedings of the 1994 IEEE International Symposium on Information Theory - Trodheim, Norw
Duration: 27 Jun 19941 Jul 1994

Conference

ConferenceProceedings of the 1994 IEEE International Symposium on Information Theory
CityTrodheim, Norw
Period27/06/941/07/94

Fingerprint

Dive into the research topics of 'Sequential sequence estimation for CPM'. Together they form a unique fingerprint.

Cite this