Universal Mobile Telecommunications System (UMTS) and High-Speed Downlink Packet Access (HSDPA) are the preferred services in many markets. However, challenges with interference and isolation among the RF cells are reducing bandwidth and coverage in dense areas and inside buildings. This article proposes a supplemental UMTS/HSDPA network architecture that cost-effectively improves service by increasing isolation to reduce interference, while increasing data speed bandwidth and coverage. The architecture improves network efficiency while reducing mobile operator production costs.

Interference and Isolation Issues
UMTS and HSDPA are noise-sensitive systems. The data speed delivered is directly related to the signal-to-noise ratio (SNR). Every cell in a UMTS/HSDPA network use the same frequency, and signals from different cells overlap, collide, and interfere with one another. This interference is due to a lack of cell isolation. This phenomenon reduces data speeds and decreases network capacity since it causes inter-cell interference in the overlapping areas (See Figure 1).

Figure 1. Frequency overlap in traditional UMTS/HSDPA networks.

The overlap issue impacts mobile operators' ability to deliver service in urban core areas and inside of buildings, where broadband users rely on service most. Subscribers opt for UMTS/HSDPA to gain the benefits of high-speed broadband. Unfortunately, the lack of isolation limits network bandwidth in densely populated areas.

Signal Attenuation
Signals at UMTS/HSDPA frequencies attenuate quickly due to their high frequency, making it a challenge to penetrate buildings. In fact, buildings present several key challenges in terms of coverage from traditional macro networks:

High penetration losses--as signals pass through building walls they attenuate very quickly, making it nearly impossible to provide adequate service within the interior areas of many buildings (See Figure 2).

Figure 2. Signal attenuation limits penetration of macro network signals inside buildings.

High power load per user--due to poor signal quality, user devices must operate at maximum power to maximize connectivity, thereby reducing mobile battery life.

A drain in overall UMTS network capacity--dense user-communities inside buildings exhaust much of the macro network's power and capacity for a given area, limiting the ability of mobile operators to serve others in the cell site vicinity.

Lack of single cell dominance and large soft handover (SHO) zones at the edges of cells--most urban areas are covered by more than one UMTS cell, so that user devices hunt from cell to cell. This limits HSDPA performance, degrades network capacity, and limits the business case for mobile operators. When covering indoor users from the macro layer, user devices may "see" more than one serving cells inside a building (See Figure 3). As a result, user connection speeds won't likely exceed 360Kbps, despite a strong ambient signal level.

Figure 3. Two adjacent cells are providing high signal levels inside the building, but the lack of isolation will give relatively slow data service.

The most effective solution for in-building coverage is an indoor distributed antenna system (DAS). An in-building DAS establishes one signal source that is far stronger than any coming from the macro network, and it can cover every area of a building (including underground facilities) with equal signal strength. However the cost of outfitting every building with an in-building DAS can be prohibitive. This is why service providers are exploring the use of micro cells, which can penetrate buildings if they are in close proximity.

However, while micro cells raise the signal used to cover outdoor or in-building areas, they don't solve the isolation problem because of overlapping areas between cells. In the overlapping areas, the data speed on HSDPA is reduced, and UMTS mobiles will load the network with soft handover (SHO) taking up resources in more cells for the same call. The key culprits here are SHO loading, which cannibalizes UMTS capacity, and signal-to-noise ratio (SNR) degradation, which reduces data speeds.