In this work, we have investigated the soft-handoff gain of IS-95 CDMA based systems in wireless local loop environment. Rician fading channels and lack of mobility lead to significantly reduced link margin gain due to soft handoff. Therefore, coverage range extension or outage probability reduction due to soft handoff is also significantly reduced. The tradeoffs associated with the variety of technical options show that soft handoff is not as desirable in wireless fixed access networks as in cellular networks.

We study here the reverse link interference of a power controlled CDMA network in wireless fixed access deployments. We show that the directionality of the subscriber antenna unit may significantly reduce the other cell interference and consequently increase the system frequency reuse. The frequency reuse gain is strongly dependent on the channel statistics as well as the beamwidth of the subscriber antenna unit. The gain achieved in frequency reuse translates directly to a capacity increase. However, soft handoff deployments lead to increased other cell interference and thus provide lower capacity gains.

Multistage trellis-coded vector quantization (MS-TCVQ) is developed as a constrained trellis source-coding technique. The performance of the two-stage TCVQ is studied for Gaussian sources. Issues of stage-by-stage design, output alphabet selection, and complexity are addressed with emphasis on selecting and partitioning the stage codebooks. For a given rate, MS-TCVQ achieves low encoding and storage complexity compared to TCVQ, and comparisons with same-dimensional multistage vector quantization indicate a 0.5-3-dB improvement in signal-to-quantization-noise ratio.

The reverse link capacity of a cellular network based on IS-95A CDMA interim standard is investigated. The results reflect various tradeoffs associated with having different rate set voice users (8 kbps and 13 kbps vocoders) simultaneously in the network. Different frame error rate choices for two rate sets lead to different performance. The tradeoffs associated with the variety of technical options show that much of the capacity can be gained back by choosing 3% frame error rate operating point for Rate Set II users.

Signal set expansion and set partitioning ideas from trellis coded modulation (TCM) have been previously applied to develop a source coding technique known as trellis coded quantization (TCQ). It has been also known that joint optimization of source and channel encoders/decoders results in substantial performance improvement. In this work, we investigate the performance of joint TCQ/TCM systems and develop various system optimization techniques. We consider channel optimized TCQ levels and asymmetric TCM symbol sets, and propose a simulated annealing based algorithm for joint optimization of TCQ levels and TCM symbols. Using these techniques signal to quantization noise ratio (SQNR) gains of 0.5-4 dB over nonoptimized systems are attained off-line. This, in turn, provides all the gain of going to a one step higher order trellis.

The error correcting capability of the unequal error protecting (UEP) Codes is described in terms of correcting errors in specific digits of a codeword even though other digits may be decoded incorrectly. We investigate the overall distortion when these codes are applied to transmission of analog sources in a communication system. Simulation and analytic results indicate the performance improvement gained by using UEP codes for various sources and source encoding structures.>