Cognitive Radio for WWANs

 
Cognitive Radio for WWANs
In this section, we take a look at the very basic functions of a cognitive radio in a WWAN
environment.
Before beginning operations, cognitive radios must obtain an estimate of the PSD of the radio
spectrum to determine which frequencies are used and which frequencies are unused. In order to
measure the spectrum in a very accurate way, a highly sensitive radio receiver will be required to
measure signals at a cell edge. Consider the example of digital TV which is also situated at the
cell edge; the received signal will be barely above the sensitivity of the digital TV receiver. For a
cognitive radio equipped with a mobile handset to be able to detect this signal, it needs to have a
radio receiver that is considerably more sensitive than the digital TV receiver. If the cognitive radio
is not capable of detecting the digital TV signal, then it will incorrectly determine that the spectrum
is unused; thereby leading to potential interference if this radio spectrum is used, that is, the signal
transmitted by the cognitive radio (with the mobile handset) will interfere with the signal that the
digital TV is trying to decode. This situation is often referred to as the “hidden node problem.”
Another example of a hidden node is shown in Figure 9.17. In this example, transmissions
between UE1 and UE2 cannot be detected by UE3 or UE4, even though UE3 and UE4 are within
signaling distance of UE1 and UE2. Therefore, UE3 and UE4 may determine that the spectrum is
unused and decide to send out a signal and potentially interfere with the signal reception at UE1 and
UE2. These examples show the necessity for highly sensitive receivers for cognitive radios.
On the other hand, if the spectrum is occupied, the cognitive radio must also be able to estimate
the interference temperature that the primary user can tolerate, that is, the transmit power level that a
cognitive device can utilize without raising the noise floor of the primary user’s device substantially
beyond a specified level. In many cases, this specified level is of the order of 0.5 dB to 1.0 dB,
but will depend heavily on the link margin available at the primary user’s receiver. The interference
temperature can be determined with at least two pieces of information: (1) an estimate of the signal
bandwidth used by the primary user, and (2) the distance between the cognitive radio and the primary
user’s device. The signal bandwidth can be used to determine the noise floor for the primary user,
while the distance can be used to determine the received signal strength seen by the primary user
as a function of transmit power used at the cognitive radio. Assuming that the noise floor for the
primary user is allowed to rise by a prespecified level, it is easy to calculate the maximum allowed
transmit power for the cognitive radio. Of course, this analysis is very basic and should be refined
if the cognitive radio can blindly classify the type of signal and corresponding data rates used by
the primary user. This extra knowledge would determine the exact sensitivity requirements for the
primary user.
Clearly, the previous discussion depends on an accurate estimate of the radio spectrum utilization
conditions in a WWAN. As the spectrum use is constantly changing, the estimate should be updated
in real time in order to ensure that the primary users are always being protected from interference.
In addition, a simple view of the radio transmission path has been assumed in the aforementioned
analysis. In a real application, however, the propagation path from the cognitive radio to the primary user might be very complicated. For example, obstacles between the two can attenuate the signal
substantially, meaning that the cognitive radio can transmit at a much higher power level than would
normally be assumed. The other possibility that the reflections of the transmitted signal may enhance
interference at the primary receiver is also valid. Consequently, the lack of information about the
transmission paths between the transmitters and receivers in a cognitive radio network can create
a great design challenge. Until now, our discussion has been primarily focused on protecting the
primary users from interference. However, a significant challenge also lies in the design of a cognitive
radio receiver. Emissions from primary users will also result in interference to cognitive radios. As
emissions from primary users can be very high, the cognitive radios must use sophisticated radio
designs, such as a multiuser detection, a MIMO system, and so on, to deal with the interference and
to ensure that the desirable signals can be reliably delivered.
Also, to take advantage of constantly changing spectrum conditions, cognitive radios need to
be adaptive in the spectrum, power levels, modulation schemes, and MAC and other upper layer
protocols that they use. This is especially challenging for mobile cellular environments in which
channel conditions can change rapidly where mobility is concerned. Thus, considerable research still
needs to be done in these areas in order to make cognitive radio a reality for WWANs.