Integration of GPS with Other Sensors and Network Assistance - Understanding GPS: Principles and Applications - page 528

9.4.2 Requirements of the FCC Mandate
The original FCC mandate allowed for two types of solutions: network-based solu-
tions to work with all legacy phones (non-GPS), and handset-based solutions (such
as assisted GPS and E-OTD) in which the user was required to have a new handset
that is augmented with the location technology. Of course, the location determina-
tion for legacy phones must be performed within the cellular infrastructure based on
TOA measurements derived using the cellular signals. Here, the FCC set accuracy
requirements to be less stringent than for future, handset-based solutions: 67% of
emergency calls needed to be located to within 100-m accuracy, while 95% of all
emergency calls required a 300-m accuracy. For handset-based solutions, the corre-
sponding accuracies were 50m and 150m, respectively. Although these numbers
appear relatively easy for a GPS receiver to meet, especially with SA deactivated, the
location determination must ideally be performed wherever a cellular phone emer-
gency call can be made, which includes indoor locations, where the GPS system was
not designed to operate.
Note that there is a degree of ambiguity in the accuracy requirement (e.g., if
95% of cellular emergency calls were made outside, and an accurate GPS solution is
provided there that satisfies the accuracy constraints, but indoor location attempts
are never successful, is the requirement met?). The FCC published guidelines [64] for
validation of accuracy and availability, but it still leaves to interpretation how to
weight or proportion outside and in-building cases for test. A possible interpretation
of the accuracy requirement across the multiple environments for which the cell
phone can make E-911 calls is the following: each environment is weighted by the
relative frequency of E-911 calls made from that environment in deriving a compos-
ite accuracy number. Hence, it is important to understand GPS signal characteristics
in the environments in which cellular phones can be expected to operate.
9.4.2.1 Characterization of Environments
A characterization of L-band signal environments was previously reported in
[65-70], which summarize data collection campaigns at 1,600 MHz in support of
satellite telephone communications link margin studies. The proximity of the test
frequency to the GPS frequency makes this research applicable. More recent GPS
in-building cumulative distribution function (CDF) fade data has been presented in
[71, 72]. In all cases, extensive radio propagation data was collected at L-band and
analyzed to characterize the shadowing, scattering, and blocking effects of trees,
cars, and buildings. Hundreds of hours of test data were collected and analyzed.
Table 9.5 lists the environments characterized in the previously mentioned refer-
ences and summarizes the 50% median fade of the signal due to the environment.
The data in the table were extracted from charts showing fade depth versus
probability charts presented in the multiple references listed.
“Heavy Urban” with the portable unit and the three in-building environments
was chosen for the basis of further calculations as the median attenuation values
were large and expected to produce reduced GPS satellite signal availability. Mobile
and in-vehicle data in an open environment were also chosen to show the trivial case
where the received signal strength is so high that fix percentage will surely be 100%
and the important case of a unit employed inside of a car.