Positioning and Location

We have just described how you can take an omnidirectional antenna (360°coverage
from a colinear antenna consisting of a number of dipoles) and produce gain in the
horizontal axis. You can also take a directional antenna and decrease the beamwidth.
Typically, directional antennas used in cellular/PCS/3G applications will have 145°,
90°, or 60° beamwidths to support three-, four-, or six-sector cell configurations. As
beamwidth reduces, gain (forward directivity gain) increases. A 1 percent beamwidth
dish antenna can deliver up to 50 dBi of gain. As beamwidth decreases, the Node
B/base station is able to discriminate the position of mobiles (angle of arrival and direction
of travel) with increasing precision.
For example, a Node B with an omnidirectional antenna knows that a mobile is
within its cell radius but does not known where the mobile is within the cell. With a
three-, four-, or six- sector cell using directional antennas, the Node B knows which
sector the mobile is in. As the mobile moves from sector to sector, direction of travel is
known. In addition, TDMA and CDMA cellular networks all have synchronous
uplinks. In TDMA networks (GSM and US TDMA) the bit delay introduced by the
round-trip between base station and handset is calculated and the mobile is timeadvanced
(the transmit slot is moved closer to the receive slot), so that all mobiles
within the serving cell arrive back at the base station in time with each other. GSM, for
example, is able to time-advance by 64-bit symbols. A bit symbol period is equivalent
to just under 1 km of flight time; therefore, the number of bit periods of timing advance
provide an indication of how far the mobile is away from the base station.
The same time synchronization process is used in IMT2000DS and CDMA2000. Synchronization
is achieved by locking the mobile to a short code burst (actually a continuous
stream of short code bursts) from the Node B. Achip symbol duration (0.26 μs) is
equivalent to 70 meters of flight time, which means potentially very accurate distance
information is available (for free) from the air interface. This is the basis for networkassisted
positioning and location services, which combine angle of arrival information
(from sectored antennas) with distance information.

A mobile can be seen by (and, in IMT2000 and CDMA2000, is supported by) more
than one base station/Node B giving additional positional information. Similarly, the
handset can see more than one base station. Because the positions (longitude and latitude)
of the base stations/Node Bs are known, then either a Node B or handset can
work out the handset’s position.
Value or quality in positioning/location depends on the accuracy of the fix, reliability
(how often the signal is usable), the consistency of the accuracy of the fix, and how
long it takes to make the fix (delay and delay variability). These in turn have an impact
on the power budget of the handset.
Table 13.2 compares the main options for network-based and handset-based location
and positioning systems. Cell ID is the simplest, but accuracy is variable, because
it is dependent on network density, as follows:
 Time difference of arrival (TDOA) uses time advance information from three
base stations and signal strength to give an accuracy of between 300 and 1100
meters. This involves a network upgrade but no handset upgrade.
 TDOA/AOA adds angle of arrival using (smart) phased array antennas to
improve accuracy—between 40 and 400 meters, but at the cost of needing to
install more complex antenna arrays.
 E-OTD (Enhanced Observed Time Difference) uses the handset to collect information
about the time of arrival of signals from the base station and at a number
of prespecified location measurement points. This is potentially more
accurate than TDOA/AOA and more accurate than TDOA but needs a handset
upgrade.
There are then three existing handset-based/satellite-based options: GPS (the U.S.
Global Positioning System using 24 satellites)—case studied briefly in Chapter 15),
GLONASS (Global Navigation Satellite System, the Russian equivalent), and, possibly
longer term, Galileo (the European equivalent to GPS). The advantage of a satellite fix
is that, provided at least four satellites can be acquired, satellite fixing gives you longitude,
latitude, and altitude.
Hybrid schemes also exist using GPS and network information. This is because GPS
works very well in applications where the handset has a clear line of site to three,
preferably four, satellites—for example, in rural areas without a lot of nearby high
buildings. In an urban environment, satellite line of site is often blocked by buildings,
but generally network density is high, so cell ID or sector ID can be used. In addition,
the network knows, more or less, where the mobile is (geographic location) and the
time, so it can tell the handset which GPS satellites are overhead—that is, which PN
sequences to run. This significantly reduces time to fix.
Each GPS satellite uses a 1.023 Mcps PN code onto which is modulated a 50 bps
navigation message consisting of the time (repeated every 6 seconds), ephemeris
(where the satellite is in orbit, repeated every 30 seconds), and an almanac (where all
the satellites are, repeated every 12.5 minutes). Assisted GPS saves the handset from
having to store or act on this almanac information. Acquisition times of 100 ms
are claimed to be achievable with A-GPS at a power budget of 200 mW per fix.

As with all positioning systems, distance is value—in this case, the smaller the distance
between the actual position of the user and the calculated position, the higher the
value. In GPS, each nanosecond of error represents one foot of measurement error. Certain
effects limit the ultimate accuracy of GPS. The density of the atmosphere changes
over time. This, and changing gravitational effects, influence the speed at which radio
waves travel, so it is impossible to realize absolute accuracy. However, differential GPS
schemes, in which signals are calibrated against known locations, can give accuracy
down to fractions of a centimeter—sufficiently accurate to detect an earthquake tremor
or detect problems with large structures (bridges, skyscrapers, and dams for example).
The practical problem of implementing GPS in a handset is the interference caused
by the phone to the GPS receiver. Either the phone needs to stop functioning while
measurements are made, which places a premium on acquisition time, or considerable
care has to be taken with handset antenna configuration and layout.
In Chapter 10, on handset hardware, we also mentioned the logic of adding a digital
compass and infrared distance measurement to the handset, to provide the capability
to identify what the handset (and hopefully the handset user) is looking at. The
handset can then do a download from the geocoded database in which Web-based
information can be searched by geographic location, longitude, latitude, and height.
This means the handset displays information on the object being pointed at.
In the United States, the E-911 directive requires network operators to be able to provide
positioning information to public safety authorities—emergency rescue services
for example. This places an additional premium on positioning and location capability. 313