Adaptive channel estimation and frequency domain channel equalization in DS-CDMA/OFDM composite radio environment for software defined radio receivers

The main contribution of this thesis is the successful development of a generic frequency domain equalizer (FDE)-based receiver architecture aiming at the application in the DS-CDMA/OFDM composite radio environment. The proposed receiver architecture featuring low-cost and good performance has great potential to be implemented via software defined radio (SDR) technology.; Firstly, we present a brief introduction to the SDR technology, including software architecture and hardware architecture. Secondly, the system models for direct sequence code division multiple access (DS-CDMA) systems and Orthogonal Frequency Division Multiplexing (OFDM) systems are introduced. Moreover, as the theoretical basis, some important wireless channel models are provided and analyzed.; Utilizing the existing channel models, a frequency domain adaptive estimation of Doppler shifts for multiple Doppler subpaths in DS-CDMA systems is proposed. By modeling the doubly selective channel using a basis expansion model (BEM), the proposed Expectation-Maximization (EM) algorithm-based estimation method can obtain accurate information of Doppler shifts. Moreover, we propose an iterative algorithm for frequency domain joint estimation of synchronization parameter and channel impulse response (CIR) in DS-CDMA systems. Based on the estimated information of Doppler shifts, synchronization parameter, and CIR, an FDE-based receiver architecture is developed to exploit Doppler diversity in the frequency domain. The proposed estimation algorithm can be easily migrated for the application in OFDM systems with minor modifications. Based on the proposed frequency domain channel estimation scheme, a generic frequency domain equalizer (FDE)-based receiver architecture is eventually proposed. This architecture can be readily implemented by the SDR technology. Our results manifest that this receiver architecture is promising and can achieve good performance with low cost.