Digital Signal Processing Reference
In-Depth Information
strategy is proposed to address both (1) and (3). Next, a technique is presented to
communicate the local estimate through the network and achieve a power control
considering not only distance, but also direction as motivated by conclusion (2).
The techniques are based on a design time model that is calibrated at run time based
on the monitoring carried out by the cognitive radio.
4.2 The Sensing Problem
Controlling or adapting spectrum use in space is typically achieved using Trans-
mit Power Control (TPC). Many techniques have been proposed to achieve optimal
power control through central coordination or in a distributed setting [50]. Each of
these techniques however requires a detailed knowledge of the propagation condi-
tions of two conflicting transmitters towards their intended receivers. When it is not
possible to rely on such channel state feedback from the receivers, this information
is typically obtained through sensing for transmitters. To detect potentially shad-
owed existing transmitters, the sensitivity requirements of secondary transmitters
needs to be set very high [51]. Even when the required sensitivity can be achieved,
the allowed power for the secondary transmitter is typically much lower than the
power that could actually be tolerated by the primary receiver since large safety
margins are required to account for unpredictable shadowing. Indeed, as illustrated
in Fig.
4.1
, the channel sensed between the primary and secondary transmitter only
reflects the channel characteristics along this specific path and safety margins need
to be introduced to avoid interference along other paths towards the receivers. An
alternative approach is the so-called geolocation database.
4.3 Distributed Distance-to-Contour Estimation
4.3.1 Algorithm Overview and Design Decisions
The problem that should be solved can be stated as follows. For each potential
secondary transmitter, this cognitive secondary transmitter should compute its al-
lowable transmission power, so that a maximum number of receiving nodes can
be reached without interfering with (potentially multiple) primary transmitters. In
the following discussion motivates that this amounts to minimizing received power
contour-to-contour distances.
On a high level the algorithm works as follows. First, nodes which sense the
primary transmitter compute a robust estimate of the local RSSI. Next, the short-
est distance to the interior nodes, i.e., nodes within the propagation contour or with
RSSI above a given threshold, is propagated through the network. This distance can
be used by a secondary transmitter to conservatively estimate an initial transmission
power. Next, the secondary transmitter iteratively estimates the minimal distance
