A new MAP channel decoder trellis path metric for a CDMA mobile subject to channel estimation errors

• Published in: 2004 IEEE Itnernational Conference on Communications Vol. 5; pp. 2879 - 2883 Vol.5
• Main Authors: Messier, G.G., Krzymien, W.A.
• Format: Conference Proceeding
• Language: English
• Published: Piscataway, New Jersey IEEE 2004
• Subjects: Applied sciences; Channel estimation; Coding, codes; Concatenated codes; Decoding; Detection, estimation, filtering, equalization, prediction; Equipments and installations; Exact sciences and technology; Fading; Frequency estimation; Information, signal and communications theory; Interference; Mobile radiocommunication systems; Multiaccess communication; Multipath channels; Probability distribution; Radiocommunications; Signal and communications theory; Signal, noise; Systems, networks and services of telecommunications; Telecommunications; Telecommunications and information theory; Transmission and modulation (techniques and equipments); Uncertainty; Performance evaluation; Mobile radiocommunication; Estimation error; Parameter estimation; Probabilistic approach; Error estimation; Wireless telecommunication; Probability distribution; Turbo code; Function approximation; Characteristic function; Simulation; Channel estimation; Code division multiple access; Analytical method; Rake receiver
• ISBN: 0780385330; 9780780385337
• DOI: 10.1109/ICC.2004.1313055

A new trellis path metric is derived for the CDMA mobile MAP and logMAP turbo decoders. This metric accounts for the non-Gaussian received symbol probability distribution seen at the output of a mobile RAKE receiver when the RAKE combining weights are subject to channel estimation error. Since an exact analytical expression for this distribution is difficult to find, the metric is instead derived by finding the characteristic function of an approximation to the distribution. Simulations are then used to compare how the CDMA forward link performs when the mobile uses either the conventional logMAP trellis path metric or the new metric. The results indicate that while the conventional metric is still the best choice when channel estimation error is small, the new metric provides a significant performance improvement when channel estimation error is large.