Digital Signal Processing Reference
In-Depth Information
terminals inside or outside the network,
...
). Increased interference results in an in-
creased packet error rate at the physical layer or an increased packet collision rate at
the MAC layer. For the user, the competition for the spectrum results in an increased
channel access delay or a lowered achieved throughput. The impact of the detailed
actions of each other terminal in the network has a huge impact on the achieved
performance and is hence the main reason why current radios should adapt, intelli-
gently.
3.3.3.3 Fairness
Wireless spectrum is a shared resource. This, however, does not imply that every-
body has the same access to it. For instance, while water is a shared resource, in
a remote village in Africa it is generally more difficult to access water than in a
European town. Unfairness can however be compensated, for instance by installing
better irrigation systems in Africa.
Similarly, some wireless devices may experience good channel conditions, while
others are experiencing worse situations. This unfairness can be compensated by
allowing users with bad channels to access the spectrum longer. Unfairness can
also be caused by a disparity of different application requirements. We have, for
instance, shown that energy-sensitive wireless devices can throttle the performance
of throughput-sensitive devices when both can access the spectrum equally [19]. Is
it fair that energy-sensitive users conserve energy, at the expense of the throughput-
sensitive users?
This question is difficult to answer as today fairness criteria
2
remain very am-
biguous and local. A strong debate is ongoing to find a unifying fairness definition.
Only when this debate closes and such a definition is found, can fairness be truly
regarded as an objective.
3.3.4 Run-Time Complexity
We have shown that the global design problem is very challenging. However, the
goal remains to design an efficient run-time procedure for a radio that can be exe-
cuted on a handheld, mobile smart or cognitive radio. Wireless communication de-
vices are typically small and battery-operated devices. Hence, algorithms designed
for these devices should be run with limited computation resources. Combining the
inherent complexity with limited capabilities is the last challenge that should be ad-
dressed. The focus of this topic is mainly on the run-time operation, that consists
of four phases as already introduced in the introduction of this chapter: observing
the RT situation, determining the scenario, executing the RT procedure and finally
2
For instance, in wireless networks a plethora of fairness indices are currently being used. We refer
interested readers to [40, 41].
