|Institution:||Curtin University of Technology|
|Keywords:||code division multiple access wireless networks, signal formats, DS CDMA wireless networks.|
|Full text PDF:||http://espace.library.curtin.edu.au:80/R/?func=dbin-jump-full&object_id=10408&local_base=gen01-era02|
One of the fundamental problems related to the development of direct sequence code division multiple access (DS CDMA) wireless data networks is design of spreading sequences possessing semi-optimal characteristics. In this thesis, we introduce three new methods to design spreading sequences, which can be optimised to achieve the desired characteristics.We show that the level of MAI for the DS CDMA systems utilising the example sets of sequences designed by the use of these techniques can be relatively low, compare to the case when the well known Gold-like sequences  are used. In addition, we show that by using one of the methods introduced in the thesis, we can construct sets of orthogonal sequences possessing acceptable correlation properties, even for an asynchronous operation, while another of the introduced methods can be used if design of sequences of an arbitrary length is required.Our new methods to design complex polyphase sequences are orientated towards the short length sequences, as a target application for them are high data rate wireless networks. Those methods are based on using discretised chirp pulses, pulses consisting of discretised multiple chirps, or linear combinations of them. In order to achieve orthogonality among the designed polyphase sequences, we combined the sequences based on superimposed chirps and double chirps with the sequences derived from the orthogonal Walsh functions.Finally, we utilise the three most promising sequence sets designed by the use of die introduced methods to simulate the multiuser DS CDMA systems. We compare performance of those simulated systems with the performance of the simulated system utilising 15-chip Gold- like sequences. The comparison results indicate that by using our design methods, we can produce useful sequence sets for applications where short spreading sequences are required. The presented ++ results also demonstrate that the performance of systems utilising those sequences can be significantly better in terms of the number of simultaneously active users or bit error rate (BER) that the performance of the system employing Gold or Gold-like sequences of the similar length.