MoPro Chapter 8.3

8.3 Slant-Path Building Penetration Measurements at L- and S-Band

In this section, we describe the results of another campaign of Vogel et al. [1995a, 1995b] who executed L- and S-Band slant path fade measurements in six different buildings employing a tower-mounted transmitter and a dual frequency receiver system. Separate circularly polarized antennas were used at the receiving end of the link. The objective of this set of measurements was to provide inside building signal loss information associated with personal communications satellite design. In particular, the correlation of fading inside buildings between frequencies near 1620 MHz and 2500 MHz was examined. The measurements described in the previous section [Vogel and Torrence, 1993] were targeted towards the application of broadcasting from geostationary satellites as they used a relatively directive receiving antenna in their measurements. In the effort described here, azimuthally omni-directional antennas were employed which interact to a greater degree with the multipath environment.

8.3.1 Experimental Description

The measurement system is comprised of a single transmitting antenna located approximately 20 m from the ground atop a tower mounted on a van. Transmissions were executed in one of two modes. Mode 1 corresponded to the case in which transmission occurred at simultaneous fixed frequencies of 1618 and 2492 MHz. Mode 2 corresponded to the case in which simultaneous radiation occurred at two swept frequency ranges, namely, 1580-1780 MHz and 2330-2530 MHz. The L- and S-Band receiving antennas were spaced 5 cm apart and both were mounted on a positioner (described in the previous section) capable of automatically scanning either in the vertical or horizontal at 5 cm increments over a range of 80 cm (16 positions). As mentioned, the receiving antennas had patterns that were azimuthally omni-directional with a peak gain at about 30° in elevation. When Mode 1 was selected, the processed outputs were relative power level versus time for L- and S-Band with a time sampling rate of 100 Hz. When Mode 2 was selected, processing resulted in frequency spectra over ± 80 MHz relative to the center frequencies of 1680 and 2430 MHz with about 1 MHz resolution and obtained over a 0.1 s sweep duration. In this mode, 160 sample pairs for L- and S-Band were saved to a computer file for every sweep duration of 0.1 s. Subsequent to the co-polarized measurements, the experimental procedures were repeated with cross-polarized receiving antennas in order to establish the efficacy of frequency reuse employing multiple polarizations.

Measurements were made into six different buildings during the Fall of 1994. The names of these buildings, pertinent construction details, and the path elevation angles are given in Table 8-7. EERL, The Commons, and the Farmhouse were previously referred to by Vogel and Torrence [1993] (Section 8.2) as Sites 1, 3, and 5, respectively in Table 8-2.

The van with processing instrumentation and mounted tower was placed outside of each of the buildings, and transmissions from the radiating antenna simulated those emanating from a satellite. Each experiment included a calibration procedure during which the co-polarized power level was measured outside the building. This co-polarized power level was used as a relative reference (clear line-of-sight) of the power measurement interior to the building.

Table 8-7: Building characteristics and measurement parameters.

Building Name	Year of Construction	Construction Type	Roof Type	Avg. El.(°)	Distance Meas. (m)
Commons	1987	concrete tilt wall	tar	16	16
EERL Office	1944	block brick	tar	30	8.8
Farmhouse	1880	wood frame	wood shingle	57	19.2
House	1958	wood frame	composition	40	12
Motel	1980	brick	composition	26	8
Store	1967	steel frame	tar	37	16

8.3.2 Stability of Measurement

An example of a 50 second time series of power is given in Figure 8-7 for L- and S-Bands for the Commons for both co-polarized and cross-polarized levels at fixed antenna positions. The standard deviation for the case shown is less than 0.2 dB at L-Band and S-Bands. Similar stability measurements were executed at the five other sites and the overall average of the standard deviations was less than 0.5 dB at both frequencies. This variability is a measure of the uncertainty in the measurements and reflects system stability, possible transmitter motion, people motion, and ambient interference.

Figure 8-7: Example of 50 s time series of relative signal level measured at the Commons for both co-polarized and cross-polarized levels and fixed antenna positions.

8.3.3 Space Variability

Figure 8-8 illustrates the spatial variation of power levels received at 1618 and 2492 MHz with co-polarized and cross-polarized antennas as a function of position inside the Commons Building, and Figure 8-9 summarizes the statistics for the median, 5%, and 95% levels for all the site locations. Table 8-8 lists the median values for the L- and S-Band frequencies as well as the overall average median at these wavelengths. Table 8-9 and Table 8-10 also summarize the mean and standard deviation for the six site locations for the co-polarized and cross-polarized cases, respectively. We note from Table 8-8 and Table 8-9 that the overall average median and mean power levels show 1.0 dB and 0.5 dB larger fades at S-Band relative to L-Band. In fact, for some of the sites the fading at L-Band is larger than at S-Band. We may conclude from these results that a small difference in the fading exists at the two frequencies.

The relative attenuation at L- and S-Bands is complicated in that the fading into buildings depends on both the absorption by building materials and the reflection and scattering properties of the building skeleton and ambient obstacles. A steel mesh, for example, may be less transmissive at the longer L-Band wavelength than at the shorter S-Band wavelength, resulting in smaller attenuation for the latter. On average, the overall mean powers (over all site locations) are -14.6 dB and -15.1 dB at L- and S-Bands, respectively, with a standard deviation of 6.5 dB at both wavelengths. It is also interesting to note that the power losses are for the most part uncorrelated, as the average correlation coefficient is 0.28.

Figure 8-8: Received power levels at 1618 MHz and 2492 MHz inside the Commons. The co-polarized and cross-polarized levels are the upper and lower figures, respectively, with the upper traces representing L-Band (left scale) and lower traces S-Band (right scale).

Figure 8-9: Summary of statistics at L- and S-Band giving median, 95%, and 5% levels of the relative power losses inside the six buildings for the case in which the antennas were moved horizontally over multiple 80 cm intervals.

Table 8-8: Summary of medians of the relative signal loss inside the six buildings at L- (1618 MHz) and S-Band (2492 MHz).

	Median Fade
Site	L-Band	S-Band
Commons	-18.6	-17.1
EERL	-14.7	-15.1
Farmhouse	-6.5	-7.5
House	-9.1	-8.4
Motel	-18.5	-19.7
Store	-13.7	-19.2
Average	-13.5	-14.5

The overall averaged cross-polarized relative power levels are only 2.3 and 1.6 dB smaller than the corresponding co-polarized power levels at L- and S-Band, respectively (Table 8-9 and Table 8-10). For example, the average relative cross-polarized power levels at L- and S-Bands are -16.9 dB and -16.7 dB, whereas the average relative co-polarized levels are -14.6 dB and -15.1 dB. These results suggest that the isolation between polarizations inside buildings is small, obviating the possibility of employing intra-system frequency re-use for gaining additional spectral capacity, although not the possibility of inter-system frequency re-use for achieving higher isolation between different satellite communications systems.

Table 8-9: Summary of statistics associated with spatial variations of L- (1618 MHz) and S-Band (2492 MHz) co-polarized relative power levels during spatial scans inside buildings.

	L-Band		S-Band
Building	Mean (dB)	STD (dB)	Mean (dB)	STD (dB)	Correlation
Commons	-19.9	8.1	-18.0	7.2	0.56
EERL	-15.0	6.9	-15.6	6.9	0.34
Farmhouse	-7.4	6.1	-7.9	7.1	0.28
House	-10.6	5.7	-9.6	5.1	0.13
Motel	-19.8	6.3	-20.1	5.6	0.05
Store	-14.7	6.2	-19.5	7.3	0.34
Average	-14.6	6.5	-15.1	6.5	0.28

Table 8-10: Summary of statistics associated with spatial variations of L- (1618 MHz) and S-Band (2492 MHz) cross-polarized relative power levels during spatial scans inside buildings.

	L-Band		S-Band
Building	Mean (dB)	STD (dB)	Mean (dB)	STD (dB)	Correlation
Commons	-16.0	6.1	-15.5	7.1	0.21
EERL	-18.1	5.4	-16.0	5.4	0.01
Farmhouse	-10.3	8.0	-11.6	8.7	0.54
House	-16.2	5.6	-15.4	5.8	0.12
Motel	-21.0	6.0	-19.8	5.1	0.16
Store	-20.0	6.4	-22.0	6.0	0.15
Average	-16.9	6.25	-16.7	6.35	0.20

Table 8-11 summarizes the statistics associated with the instantaneous differences in the S-and L-Band powers received during the spatial scans at the various site locations. The difference power level was found to be Gaussian distributed as depicted by the example given in Figure 8-10. The mean difference indicates whether on average there is a bias of the power levels at one frequency relative to the other. We note that on average, the bias ranges from +1.9 dB for the Commons to -4.8 dB for the Store sites, with an overall average bias of approximately -0.6 dB. For all sites, the standard deviations are significantly larger than the respective means. The relative lack of correlation between the instantaneous L- and S-Band fades suggests the difficulty of applying power control to a link operating in this environment.

Table 8-11: Statistics associated with differences between S-Band and L-Band co-polarized powers for spatial scans at each of the site locations.

Building	Mean Difference (dB)	STD (dB)
Commons	1.9	7.2
EERL	-0.6	7.9
Farmhouse	-0.5	7.9
House	1.0	7.2
Motel	-0.30	8.2
Store	-4.8	7.8
Average	-0.55	7.7

Figure 8-10: Cumulative probability distribution of signal level difference of the instantaneous S-Band minus the L-Band powers. The indicated straight line implies a Gaussian distribution with a mean and standard deviation of 1.9 dB and 7.2 dB, respectively.

8.3.4 Frequency Variability

Rapid changes of the power with frequency occur only when the average power level is comparable to the diffusely scattered power. This concept is illustrated in the frequency scan example in Figure 8-11 which shows the power levels relative to the center frequencies of 1680 MHz and 2430 MHz obtained at the EERL Building. Four cases were selected from all positions, namely frequency scans with mean relative received power levels of -20, -15, -10, and -5 dB. It is noted that when the mean signal level is -5 dB, the variability of the signal level over the 160 MHz band is relatively small (e.g., -3 dB maximum and -10 dB minimum for both L- and S-Bands). On the other hand, the -20 dB mean level shows signal loss deviations that are significantly larger over the frequency scan (e.g., -12 dB maximum and -38 dB minimum at L-Band).

Figure 8-11: Relative power loss variation with frequency at L-Band (upper figure) and S-Band (lower figure) for various mean power levels over the 160 MHz bandwidth.

An indicator of the frequency variability is the fade slope as a function of frequency defined by

(8-6)

where dS is the change in the received co-polarized power over the 1 MHz frequency interval df. A regression expression has been established relating the standard deviation of the fade slope to the mean signal level m_SLover the frequency span for each of the buildings at both L- and S-Band. This expression is given by given by

(8-7)

where a, b, c are tabulated in Table 8-12 for the L- and S-Band wavelengths. Summary statistics describing the mean and standard deviation of the fade slopes at the various site locations are given in Figure 8-12. It is noted from Figure 8-12 that the median fade slopes (upper figure) varies between 0.5 and 1.3 dB/MHz and the corresponding median standard deviations (lower figure) vary between 1.8 and 0.7 dB/MHz. We conclude from these levels that for a resolution of 1 MHz the frequency distortion is expected to be small over the frequency range 1680 to 2430 MHz.

Table 8-12: Tabulation of standard deviation parameters in (8-7) for different site locations at L- and S-Band.

	L-Band			S-Band
Location	a	b	c	a	b	c
Commons	-0.589	-0.11	0.000	-0.52	-0.123	-0.001
EERL	0.323	0.038	0.004	0.296	0.027	0.003
Farm	0.298	-0.047	0.003	0.196	-0.014	0.003
House	0.169	-0.015	0.002	0.237	0.026	0.004
Motel	0.531	0.063	0.004	0.342	0.047	0.004
Store	0.379	0.008	0.004	0.549	0.057	0.004

Figure 8-12: Summary statistics giving the median, 95% and 5% levels of the mean fade slope (upper figure) and the standard deviation of the fade slope (lower figure) at the various site locations.

Home

Forward