A Picture is Worth 1,000 Words

Signal-strength coverage remains the primary parameter by which wireless networks are judged. Quality-of-service metrics such as automated voice quality mean opinion scoring has emerged, providing a better overall measure of wireless network performance. However, measuring signal strength continues to be the carriers' preferred medium.

One of the main problems in measuring signal strength is getting a true picture of the coverage. Field strength decreases as you move away from the transmitting site. However, the measurements will vary considerably during the drive, and the signal will continue to vary over time even when stationary. Therefore, measured field strength is dependent on location, environment and time.

Besides environmental changes, received signal strength varies due to multipath, which is caused by signals arriving at the receiving antenna at different times and angles from the transmitter because they have been scattered along the way. Each of these signal paths differs in length (and therefore in time and phase). When they combine at the receiver, the separate paths can either meet in phase and add or meet out of phase and cancel. This adding and canceling causes large fluctuations in the received signal strength.

Two main types of fading occur: Rayleigh fading, caused by multipath, and Log normal fading, which is produced by shadowing or signal defraction from obstructions. Although Rice and Doppler are other fading characteristics, Log normal and Rayleigh receive the most attention and sometimes are referred to simply as long-term and short-term fading, respectively.

SCANNING METHODS Measuring such a changing environment accurately depends on the scanning method.

Carriers typically use prediction tools to generate the majority of network coverage plots for engineering planning requirements. Predictive plots show full coverage without the need for actual signal-strength surveys. This makes them ideal for network planning and for showing coverage beyond normal measured areas. However, prediction tools need to be calibrated to give the best possible results; therefore, accurate measured data is essential for calibration purposes.

Drive testing normally uses either specialized CW receivers or scanning features available within some mobile stations. Specialized receivers are more accurate, typically measuring +/-1dB. Scanning features in test mobiles will give only the accuracy of a normal phone, up to +/-6dB. Also, specialized scanners usually have variable scan rates and selectable channels, whereas mobile stations normally offer fixed channel lists at a fixed scan rate. Signal-strength measurements can be taken off the network or from a test transmitter for proving new site locations or in-building surveys. However, it is important to choose the correct IF bandwidth in the receiving equipment to match the transmitter to get accurate measurements.

Simple, instantaneous measurements will give a confused picture of signal-strength coverage. Because of this, averaging techniques were developed to show a better picture of the actual coverage. Averaging techniques should be small enough for Log normal fading to be constant and large enough to remove the effect of Rayleigh fading.

One of the most popular techniques is known as the Lee Average. In his Estimate of Local Average Power of a Mobile Radio Signal, William Lee says that an averaging method should consist of collecting and averaging at least 36 individual samples uniformly spread over 20 to 40 wavelengths. This should ensure measurement accuracy within 1.17dB of the true average local value. In practice, most engineers use 36 samples over 40 wavelengths. You should carry out the averaging during the collection process because it tends to be more accurate than applying averaging techniques during the post-processing phase. Other averaging techniques exist, of course, but this is one of the most reliable.

THE RIGHT TOOLS To measure at the sample rate mentioned above requires a receiver that has adequate measurement speed and, if multiple frequencies are to be measured simultaneously, sufficient switching and settling time between frequencies. The navigation equipment used to calculate the correct intervals between measurements must be fast enough to achieve accurate results. Thirty-six equally spaced measurements in every 40 wavelengths at safe driving speed normally will mean sub-meter measurements, typically better than standard navigation devices such as GPS reporting at once per second. The higher the frequency being measured and the more simultaneous frequencies, the better the equipment will need to be to achieve reliable results. The alternative is single-frequency scanning at a snail's pace.

High-quality signal strength measurements are the cornerstone of wireless network planning. They provide accurate pictures of how coverage varies in different environments, such as rural, suburban, urban and in-building, with different antenna configurations and mounting arrangements. Knowing the actual network coverage helps carriers to avoid co-channel interference problems and areas of poor coverage. It also ensures that you have enough signal strength to stop problems in high-fading environments.

Ogley is Safco Technologies director of business development. His e-mail address is dgo@safco.com.