Cellular Systems

Wireless communications became a commercial success in the early 1980s with the introduction of cellular communication systems in the United States, Japan, and Europe. All of these systems were based on frequency division multiple-access whereby a user during a call was assigned a given frequency for transmission to a base station (uplink) and a given frequency for reception from the base station (downlink). The modulation technique adopted was frequency modulation (FM) for the voice signal. The bandwidth assigned to each user was around 30 kHz (depending on the country). Frequency division multiple access (FDMA) and FM modulation were well known techniques/technology available to system designers at the time. It is interesting to note that a university professor in the early 1980s actually proposed code division multiple-access (CDMA) as a technique for cellular communications, but AT&T/Bell Labs (pre-divestiture) discredited it. At the time digital technology was not ready for a modulation technique requiring sophisticated processing algorithms. Nevertheless, the basic research component of this effort affected future wireless communication systems.

The designers of cellular systems faced many challenges that were unique to mobile environments. These challenges included time-varying multipath fading and interference from other users. The problem of multipath fading was handled by simply increasing the amount of transmitted power. Most of the cell phones were mounted in cars, handheld units were not available, and battery power was not the initial driving force. Interference was minimized by not reusing a given frequency in an adjacent cell. Finally, the application was purely voice communications.

In Europe, many different first generation systems operated in different frequency bands, and thus roaming between countries was not possible with a single cell phone. This lack of interoperability in Europe was one of the key drivers towards developing a second generation system in Europe. In the United States, the increased use of cell phones in the mid-1980s led to a shortage of capacity in major markets (Los Angeles, New York, Chicago). In order to achieve higher capacity (users/cell/Hz) second generation systems began development in the mid-late 1980s. Second generation systems (in the United States, Europe, and Japan) are all based on digital transmission techniques whereby the voice signal is encoded into a sequence of bits. The voice-coded data is then encoded for error correction and modulated using digital modulation techniques. Going to a digital modulation technique has several advantages, including the spectral efficiency of digital modulation, the capability of combining speech and data services, and the improved security of digital techniques. The first proposal for second generation systems in the United States was based on time-division multiple-access (TDMA) whereby each 30 kHz of bandwidth was slotted in time so that three users could use the same spectrum. This was possible because of the compression algorithms applied to speech waveforms. In these systems error correction was introduced to mitigate the effect of multipath fading. Interference is not a predominant issue in this design as different users use either different frequency or time slots or nonadjacent cells. Like first generation systems, essentially the only application for second generation systems (both cellular and personal communication systems (PCS) that use the 1900 MHz frequency band) is voice, although some low speed data communication capabilities are also possible.

It is interesting to note that the modulation technique chosen in Europe (Gaussian minimum shift-keying) for second generation systems is a constant envelope technique while the modulation technique chosen in the United States and Japan is pi/4 differential phase shift keying (DPSK) with raised cosine filtering, which is a nonconstant envelope technique. Constant envelope techniques ensure that the envelope of the transmitted signal is a constant. This fact allows the power amplifier used by the mobile system to operate near saturation without distorting the signal. The most energy efficient operation of an amplifier is near saturation as this is when the power added efficiency is largest. The disadvantage of constant envelope modulation techniques is that their bandwidth efficiency is small relative to modulation techniques that have fluctuating envelopes. On the other hand nonconstant envelope modulation techniques need very linear amplification in order to preserve the signal shape (no distortion). Nonlinearities applied to a nonconstant envelope technique create both adjacent channel power and in-band intermodulation distortion. The adjacent channel power essentially widens the bandwidth occupancy of the signal while the in-band intermodulation distortion acts like additional noise in the system. No distortion is incurred when linear amplification is used with nonconstant envelope modulation techniques. However, linear amplification is possible mainly by operating the amplifier with a small input signal (large backoff) where the energy efficiency of the amplifier is smallest. This characteristic of nonlinear amplifiers makes large power efficiency and bandwidth efficiency hard to achieve simultaneously.

In the late 1980s Qualcomm made a proposal to use direct-sequence spread-spectrum multiple-access (also called code division multiple-access) for cellular systems. Qualcomm's initial claims of significant (more than 10 times) increase of capacity of cellular systems captured the attention of service providers. This effort developed into a later second generation standard known as IS-95. The increase in capacity, it was claimed, was due to exploitation of the voice activity factor, Rake reception, which allowed exploitation of the multipath fading and frequency reuse of 1 (every cell uses all frequencies), which allowed more efficient use of the frequency spectrum over a geographical area. In addition, multiple users spreading their signals over a wide bandwidth with user unique codes allowed for multiple users using the same spectrum at the same time. The IS-95 system uses 1.25 MHz of spectrum. The IS-95 system first became available commercially around 1995.

A significant event for wireless communication systems occurred in the mid-1990s when the Federal Communication Commission (FCC) decided to auction off spectra in the 1.9 GHz band. This opened up a pair 60 MHz frequency bands between the 1850-1910 and 1930-1990 MHz spectra for use by those with winning bids in the auction. This band was called the personal communication system (PCS) band, and systems operating in this band are called PCS systems. The modulation and coding techniques that were chosen for these systems were virtually all-cellular type systems except they were shifted up in frequency. The operating characteristics (coding, modulation, and multiple-access) were identical. Nevertheless this provided additional capacity for wireless communications and allowed for increased competition.

The overall goal of first and second generation systems is primarily voice communications. The innovations between first and second generations were mainly in going to digital modulation techniques and the use of error control coding. Third generation systems are now being designed and implemented to handle not only voice but high speed (from 384 kbps to 2 Mbps) data, although no one really knows what the market for these services will bear.

Other Wireless Systems

Cellular communication systems are clearly the most widely used wireless communication systems. However, other systems have different characteristics that deserve mentioning. To begin with, many military communication systems need to operate in a very different type of environment. First, in a military communication system, there may not be the possibility of dividing a region into cells and deploying a base station in each cell. Military systems tend to be highly mobile and may possibly operate in unfriendly environments where the cellular type of deployment is not possible. Military systems must also face the possibility of hostile interference (jamming). One of the consequences of not having base stations is that relaying messages is required. In other words, multihop communications over multiple wireless links is necessary. This creates a whole new class of communication problems and constraints.

Another wireless communication system of interest is a wireless local area network (WLAN), whereby a wireless link replaces the wired LAN. Currently such systems tend to operate in the Industrial-Scientific-Medical (ISM) band (902-928 MHz, 2400-2483 MHz, 5725-5780 MHz). This band is available for unlicensed use (in the United States) provided either the power levels are small enough or spread-spectrum modulation techniques are employed with larger transmitted power. The 2.4 GHz band is available world-wide while the 900 MHz band is not available is some parts of the world (e.g., Europe).

Another wireless system gaining significant attention is a cable replacement wireless system called Bluetooth. This is being developed by a consortium of communications and computer companies (e.g., Ericsson, Nokia, IBM, Intel, Motorola). The objective of a Bluetooth system is to replace the cables connecting different components in a computer system with wireless links. Because it is a cable replacement the objective is for a very low cost, low range system. It is also viewed as a means for connecting a mobile computer with a cell phone in a wireless manner such that communication between the mobile computer and a hidden (e.g., in the briefcase) cell phone can be used to connect to the Internet. It can also connect a headset to a cell phone without wires. The range is about 10 meters with 0 dBm (1 milliwatt) transmitted power, but can be increased to 100 meters with larger power. Bluetooth uses the 2.4 GHz ISM band with frequency-hopped spread-spectrum. The maximum data rate is 721 kbps. It handles up to 8 devices in a so-called "pico-net" and can have up to 10 pico-nets operating in a coverage area. The networks created with Bluetooth are ad-hoc networks implying multiple-hops per end-to-end connection.

A competing system is HomeRF. HomeRF is geared more towards higher data rates and higher transmitted power. As with Bluetooth, HomeRF is a frequency-hopping system operating in the 2.4 GHz band. Products with data rates in the range of 2 Mbps with operating distances on the order of 150 feet are possible.

The development of a variety of communication systems is shown in Fig. 2.1 as a function of data rates and user mobility/cell sizes. This figure illustrates the fact that higher data rates are possible at lower mobilities or decreased cell size. This is due to two considerations. The first (and main) consideration is that at smaller distances the propagation loss is less and thus for a given power level, higher values of Eb/N0, the received signal-to-noise ratio, are possible. Another effect is that at high mobilities the channel is harder to estimate and thus proper demodulation/decoding becomes more difficult. This is compensated for, to a certain extent, by the fact that generally error control coding works better (for a fixed block length) when the channel is memory-less (independent fading on different symbols).


Fig. 2.1. The development of a variety of communication systems is shown as a function of data rates and user mobility/cell sizes.

Published: July 2000; WTEC Hyper-Librarian