Digital Signal Processing Reference
In-Depth Information
3.4.3.1 Observe the RT Situation
①
First, the environment is observed with the monitors, defined at DT. In general, the
information to monitor and to be taken into account by the definition process of the
smart radio can also be calibrated or partially learned at DT. This is however not
considered in this topic.
The monitoring of the environment is very important in the context of Oppor-
tunistic Spectrum Access, which was probably the first well known example of the
cognitive radio paradigm. In this use case, smart radios are called secondary users
and have to closely monitor the use of the spectrum by primary users. If a primary
user is present in a certain area, the spectrum cannot be used by the secondary user.
This simple monitor and act, although the simplest instantiation of the framework
introduced in this topic, is one of the most known cases of cognitive radio.
3.4.3.2 Map RT Situation to System Scenario
②
The encountered scenario is then mapped to a specific system scenario, based on the
information derived from the monitors. The observed system scenario gives input to
the RT procedure, that is executed and possibly calibrated or even partially learned
even at DT. Potentially, the selection of system scenarios can also be calibrated or
learned. This is however not considered in this topic as we only calibrate the RT
procedure (see Fig.
3.2
).
A given scenario is often only determined uniquely when monitors across layers
and possibly even different nodes are combined. In this topic, this will be illustrated
for the case of an IEEE 802.15.4 network coexisting with an IEEE 802.11 network.
For the IEEE 802.15.4 network to find it's globally optimal channel, it is needed to
monitor both the physical layer interference and the routing parameters.
3.4.3.3 Execute RT Procedure
③
The RT procedure determines the action to be taken in each of the given scenarios.
An action can be an absolute setting of a knob to a value, or a relative increase of the
knob setting. For instance, it is possible to give an output power value for the PA, or
rather an action can be specified as increasing the output power of the PA with 2 dB.
The actions that should be taken by the smart radio can be fully determined at
DT, or only partially specified. In the latter case, the RT procedure is calibrated or
learned at RT by means of feedback.
In its most generic version, the RT procedure is hence a function of the current
environment, the observed feedback history and the DT procedure.
A RT procedure that is fully characterized at DT is illustrated in Chap. 6.The
considered use case is an IEEE 802.11 network where multiple users have to share
the medium. By means of scheduling, the power-performance of the entire network
is optimized, based on a RT procedure fully specified at DT.
