Bearing estimation techniques for improved performance spread spectrum receivers
The main topic of this thesis is the use of bearing estimation techniques combined with multiple antenna elements for spread spectrum receivers. The motivation behind this work is twofold: firstly, this type of receiver structure may offer the ability to locate the position of a mobile radio in an urban environment. Secondly, these algorithms permit the application of space division multiple access (SDMA) to cellular mobile radio, which can offer large system capacity increases. The structure of these receivers may naturally be divided into two parts: signal detection and spatial filtering blocks. The signal detection problem involves locating the bearings of the multipath components which arise from the transmission of the desired user’s signal. There are a number of approaches to this problem, but here the MUSIC algorithm will be adopted. This algorithm requires an initial estimate of the number of signals impinging on the receiver, a task which can be performed by model order determination techniques. A major deficiency of MUSIC is its inability to resolve the highly–correlated and coherent multipath signals which frequently occur in a spread spectrum system. One of the simplest ways to overcome this problem is to employ spatial smoothing techniques, which trade the size of the antenna array for the ability to resolve coherent signals. The minimum description length (MDL) is one method for determining the signal model order and it can easily be extended to calculating the required degree of spatial smoothing. In this thesis, an approach to analysing the probability of correct model order determination for the MDL with spatial smoothing is presented. The performance of MUSIC, combined with spatial smoothing, is also of great significance. Two smoothing algorithms, spatial smoothing and forward–backward spatial smoothing, are analysed to compare their performance. If SDMA techniques are to be deployed in cellular systems, it is important to first estimate the performance improvements available from applying antenna array spatial filters. Initially, an additive white Gaussian noise channel is used for estimating the capacity of a perfect power–controlled code division multiple access system with SDMA techniques. Results suggest that the mean interference levels are almost halved as the antenna array size doubles, permitting large capacity increases. More realistic multipath models for urban cellular radio channels are also considered. If the transmitter gives rise to a number of point source multipath components, the bearing estimation receiver is able to capture the signal energy of each multipath. However, when a multipath component has significant angular spread, bearing estimation receivers need to combine separate directional components, at an increased cost in complexity, to obtain similar results to a matched filter. Finally, a source location algorithm for urban environments is presented, based on bearing estimation of multipath components. This algorithm requires accurate knowledge of the positions of the major multipath reflectors present in the environment. With this knowledge it is possible to determine the position of a transmitting mobile unit. Simulation results suggest that the algorithm is very sensitive to angular separation of the multipath components used for the source location technique.