Since lower bandwidth implementations of the Long-Term Evolution (LTE) 4G technology (1.4 and 3 MHz) are significantly less efficient, providers of wireless broadband rural access in the spectrally narrow 450 MHz band face the question of whether migrating to LTE can actually provide the desired and promised leap in performance, and whether existing and in many cases deployed 3G technologies such as 3G CDMA Evolution Data-Optimized (EVDO) may continue to offer comparable benefits. This work addresses this question via numerical simulations for a set of conditions aimed at defining a level playing field for fairly comparing these two technologies in terms of up and downlink cell aggregated throughput and range.

Provision of voice services over 3G W-CDMA radio access networks is generally designed as a function of the air interface requirements of circuit-switched (CS) connections. On the other hand, packetized VoIP speech services such as Skype, which is becoming increasingly popular among 3G users, impose a more demanding set of requirements on the radio interface, carrying a penalty in terms of coverage. This work quantifies such penalty as a function of key parameters of the radio uplink. Results show that Skype coverage is always a fraction of conventional CS speech connections, and that such fraction is mainly a function of the vocoder type selected by Skype and the uplink Block Error Rate. It is found that Skype cell coverage probability ranges from 67% to 98% of the CS voice service coverage. As a result of this analysis, a set of recommendations aimed at closing the gap between CS voice and Skype coverage is formulated.

A simple mathematical model for predicting the probability that a street-level location within a macrocell enjoys full visibility to its serving base station antenna is proposed and evaluated. The single-parameter, two-tier linear model is based on processing and modeling actual antenna observation data taken at street-level locations around urban cell sites in the city of Caracas. When the only parameter in the model is allowed to take its optimal value, the correlation between data and model is 0.89. Furthermore, when this parameter is expressed as a function of the antenna height, such correlation drops only slightly to 0.87, which leads to postulating a simpler version of the model. Finally, the antenna height-dependent model leads to a closed mathematical expression for the probability of LOS for street-level locations in hexagonal cells that depends only on the ratio of antenna height to cell range.

As WiFi networks and IP telephony grow increasingly popular, the standard Distributed Control Function (DCF) MAC protocol in IEEE 802.11b, which is not designed to support realtime applications, has become a major hurdle in the effort to increase Voice over IP (VoIP) capacity. This work identifies the FAIR algorithm as a viable solution for increasing VoIP capacity in 802.11b networks, since it is aimed at providing a more balanced use of medium resources in infrastructure-based networks and it can be implemented by simply changing the MAC protocol at the Access Point only. Simulations using real traffic scenarios (including both VoIP and HTTP users) show VoIP capacity gains of 50% and more, and also provide an insight on the issues that may further accentuate the effectiveness of FAIR.

The impact of antenna downtilt on the forward links of overlayed CDMA2000 1XRTT and EV-DO systems is examined via numerical simulations. A system model for colocated IX / EV-DO networks is proposed and employed to provide a perspective on the issues and trade-off situations derived from using the same cell site antennas and, consequently, the same downtilt angles for the two networks. A set of recommendations regarding the fade margins of both downlink budgets and the optimum antenna downtilt angles as a function of the IX cell load is drawn from the simulations results.

A forward error correction (FEC) strategy and a medium access control (MAC) protocol that are thoroughly tailored to complement and support a high-speed asymmetric physical layer design based on equalization and precoding is presented and fully discussed. Both proposals exhibit a high degree of adaptability and flexibility, which allows for increased data throughput while providing a wide range of quality-of-service requirements. Fast link layer adaptation is made possible through the joint design of link and physical layers. The adaptive FEC algorithm is based on the use of variable-rate trellis coded modulation with fast channel estimation, while the MAC protocol employs a centralized, dynamic slot allocation technique. The overall system design is shown to achieve high spectral efficiency, while minimizing energy consumption at the portable unit.

The problem of identifying the optimum adaptive channel code rate that achieves a target error performance in a wireless precoded link operating over slowly varying channels is addressed. It is found that pragmatic trellis coded modulation (PTCM) with adaptive QAM constellation is a natural fit to the problem at hand. It supports variable coding rates while providing decoder simplicity. Moreover, it enables the efficient exploitation of the unique features of a precoded system. The article also describes a simple technique for fast code rate adaptation based on channel estimation algorithms that combine the use of intra-frame and inter-frame physical layer information. As a consequence of the implementation of these adaptive techniques, a significant increase in data throughput is achieved, as verified in simulations.

A novel set of techniques for microscopic space diversity in high-speed wireless systems that employ decision feedback equalization (DFE) to mitigate ISI is postulated and evaluated via simulations. Adaptation on a packet basis rather than on a symbol basis is shown to be a simple and effective choice in indoor, low-mobility environments. Antenna selection based on slicer signal-to-noise ratio and largest-ray antenna combining are shown to improve the performance of a DFE-based link by margins that are similar to those obtained by conventional selection and combining techniques operating over flat fading channels. Furthermore, in the context of asymmetric systems that employ DFE on the reverse link and channel precoding on the forward link, the proposed techniques positively impact both links without adding any circuitry to the downlink receiver.

High data rates and portability are conflicting requirements in the design of reliable indoor wireless data communication systems. Asymmetric system design addresses this problem by exploiting the major differences in the availability of resources (i.e., energy, space) that exist between base stations and portable units. Such an approach seeks to concentrate most of the signal processing tasks involved in the operation of the two-way wireless link at the base station. This paper presents and discusses the implications of a set of techniques for asymmetric physical layer system design based on the use of channel precoding for forward transmission. System performance is assessed via simulations using a realistic time-varying channel model. Simplified two-way antenna diversity implemented exclusively at the base station and the key issue of automatic gain control at the receive side of the precoded link are also addressed and evaluated. Simulation results confirm that the adoption of these techniques enables reliable digital communication at a data rate of 20 Mbit/s on both links while significantly reducing the power consumption of the portable unit.