Efficient Radio Resource Allocation Schemes and Code Optimizations for High Speed Downlink Packet Access Transmission

Abstract

An important enhancement on the Wideband Code Division Multiple Access (WCDMA) air interface of the 3G mobile communications, High Speed Downlink Packet Access (HSDPA) standard has been launched to realize higher spectral utilization efficiency. It introduces the features of multicode CDMA transmission and Adaptive Modulation and Coding (AMC) technique, which makes radio resource allocation feasible and essential. This thesis studies channel-aware resource allocation schemes, coupled with fast power adjustment and spreading code optimization techniques, for the HSDPA standard operating over frequency selective channel. A two-group resource allocation scheme is developed in order to achieve a promising balance between performance enhancement and time efficiency. It only requires calculating two parameters to specify the allocations of discrete bit rates and transmitted symbol energies in all channels. The thesis develops the calculation methods of the two parameters for interference-free and interference-present channels, respectively. For the interference-present channels, the performance of two-group allocation can be further enhanced by applying a clustering-based channel removal scheme. In order to make the two-group approach more time-efficient, reduction in matrix inversions in optimum energy calculation is then discussed. When the Minimum Mean Square Error (MMSE) equalizer is applied, optimum energy allocation can be calculated by iterating a set of eigenvalues and eigenvectors. By using the MMSE Successive Interference Cancellation (SIC) receiver, the optimum energies are calculated recursively combined with an optimum channel ordering scheme for enhancement in both system performance and time efficiency. This thesis then studies the signature optimization methods with multipath channel and examines their system performances when combined with different resource allocation methods. Two multipath-aware signature optimization methods are developed by applying iterative optimization techniques, for the system using MMSE equalizer and MMSE precoder respectively. A PAM system using complex signature sequences is also examined for improving resource utilization efficiency, where two receiving schemes are proposed to fully take advantage of PAM features. In addition by applying a short chip sampling window, a Singular Value Decomposition (SVD) based interference-free signature design method is presented.