Space-time coding and receiver design for unknown time-varying wireless channels.

摘要

This thesis considers a joint channel estimation and data detection algorithm for Multiple Space-Time Trellis Codes (MSTTCs) over high Doppler fading channels. This algorithm belongs to the Per-Survivor Processing (PSP) family, where a survivor associated with a trellis state is used for channel estimation. We propose a smoothed data detection technique to increase the probability of the survivor being a correct data sequence. A symbol by symbol Maximum A Posteriori Probability (MAP) criterion, with a fixed delay D, is employed for data detection, combined with a Kalman Predictor (KP) for channel estimation. This novel algorithm, called Smoothed Data Detection and Kalman Estimation (SDD-KE), provides a significant performance gain when D > 0 over D = 0, with a linear increase in complexity when D increases. Comparison with the Delayed Mixture Kalman Filtering (DMKF) technique shows that the SDD-KE algorithm provides important performance and complexity advantages.;This thesis also develops a joint channel estimation and data detection algorithm for unknown Doppler shift channels, called Doppler Adaptive SDD-KE (DA-SDD-KE) algorithm. This algorithm jointly detects the data, and estimates the channel as well as the Doppler shift, which can be time-variant, by incorporating an Adaptive Kalman Predictor (AKP) into the SDD-KE algorithm. Due to the lack of a model for generating a Rayleigh fading process with time-varying Doppler, this thesis proposes such a model. Simulation results show that the DA-SDD-KE algorithm performs well over fading channels with constant high normalized Dopplers of fdT = 0.03, 0.05 and over fading channels with linear time-varying Dopplers in the range of 0.01 and 0.05 (changing within a frame length of less than 300 channel symbols). Moreover, the results for quadratic Doppler functions in the ranges of 0.03 to 0.05 and 0.01 to 0.03 also show that the DA-SDD-KE algorithm performs considerably well, especially when D = 2 is used.;Motivated by the fact that all joint channel estimation and data detection schemes suffer from the phase ambiguity problem, we developed a design method that can transform any phase ambiguity prone STTC into a Phase Ambiguity Diminishing STTC (PAD-STTC). A PAD-STTC can be considered as an MSTTC with special constellation mappings. The design criteria for these mappings are developed by trading among the PAD property, diversity, and coding gain. Two PAD-STTC structures termed PAD1-STTC and PAD2-STTC are proposed, where PAD2-STTC increases the number of states over the constituent STTC. Computer simulation results indicate that PAD-STTCs can solve the phase ambiguity problem and provide a larger diversity gain than the constituent STTC over high Doppler fading channels. Moreover, with the same constituent STTC, the PAD2-STTC outperforms the PAD1-STTC at the expense of increased complexity.