| A foundation for the generation of spreading codes based upon multirate filter banks is introduced. This new class of spreading codes is proposed as an alternative to the m-sequences which are the traditional choice for a direct sequence spread spectrum system. Unlike m-sequences, the new codes are not limited to being binary valued and their construction is based upon specific known channel characteristics.; After describing the generation of the codes from a given multirate filter bank, several important properties and characteristics of the codes are formulated in terms of the subband tree filters. A methodology for designing the filters is then presented. The filter tap weights are generated by solving a multivariable constrained optimization problem. The optimization problem incorporates the properties of the codes as well as channel characteristics. Objective functions and constraints are presented for three interference channels common to a mobile communications environment: stationary interference, multiuser interference, and multipath interference.; Expressions are then developed for analyzing the bit error rate performance of the spreading codes under various channel conditions. These expressions can be used to determine the capacity, in terms of number of users in the channel, of a system using the newly developed spreading codes. Filter banks are then designed for the various channel types. Designs for the multiuser channel include synchronous and asynchronous multiuser systems, as well as, a system with partial synchronization.; Comparisons are then made with the m-sequences and the Gold codes. It is shown that the newly developed spreading codes outperform the m-sequences and Gold codes when used in channels for which the new codes were designed. It is shown that two limiting cases of spreading codes based upon multirate subband filter banks result in FDMA and TDMA systems and the remainder of the class of codes provides a tradeoff of performance in various interference scenarios. |
