The correlation between signals received by two mobile radio base-station antennas is investigated to determine spacing requirements for space diversity. Measurements of the fading of UHF signals received by two base-station horn antennas oriented at different angles with respect to the incoming mobile radio signal were made for different antenna spacings. The experimental results are compared with an analytical expression derived in this paper; they agree fairly well. A further experiment was made after removing the possible local scatterers surrounding the base station. Comparing these two experimental results, we find that the following are true. 1) Propagation in the direction of a line connecting the two base-station antennas is the critical case and requires a large separation of 70 wavelengths. As soon as the incoming wave is 10° away from the inline axis, the spacing requirement drops to 30 wavelengths. 2) Local scatterers at the base station tend to decrease the correlation between signals received at the two antennas. We conclude that an upper limit to the spacing of antennas used for diversity can be obtained and that it is within the achievable range.

Field measurements of mobile radio signals have shown that signal arrival is concentrated in elevation angles lower than 16°. This suggests use of an omnidirectional antenna with vertical directivity, instead of the usual whip antenna, to increase average signal strength as a further improvement in the advantages of diversity reception.

The effect of mutual coupling on a space diversity combined signal from an optimum resistive load network lessens when the antenna spacingdand the number of elementsMincrease. Whendis greater than 0.2λ, the cumulative distribution curves of anM-branch signal with mutual effects included show only a small difference compared to an array with no coupling, and are close to the cumulative distribution curve of an array withMindependent branches. For the average SNR of an array with 100 branches obtained from an optimum resistive load network to be within -3 dB of the SNR from independent branches the required antenna spacing is at least 0.25λ, for an in-line array and 0.5λ, for a planar array. Since a planar array is much smaller than an in-line array for equal performance, the planar array is a more desirable configuration.

Conventional space diversity reception at typical elevated base locations requites separation of 30λ for broadside incidence and even more for in-line incidence and is therefore difficult to implement. A polarization diversity system for mobile radio is proposed. This is a two-branch receiver diversity system with the advantage that the base station antennas can be spaced as closely as desired. An experimental program has been carried out to obtain the statistical properties of vertically and horizontally polarized electromagnetic waves in a suburban environment at 836 MHz. It was observed that signals of both polarizations were Rayleigh plus log normal, where one is uncorrelated and other is correlated irrespective of base or mobile antenna spacings. The local means of the two signals were highly correlated and were with ± 3 dB for almost 90 percent of the time. Variation of base transmitter heights appeared to have little effect on the ratio of the local means of the two signals. The analysis and experiment demonstrated the feasibility of providing two diversity branches at UHF by polarization diversity.