Digital Signal Processing Reference
In-Depth Information
Fig. 6.3
802.11a OFDM direct conversion transceiver
These cost, quality and resource dimensions can be layered, to facilitate run-time
information flow and ease of modeling. Finally, the mapping of the control dimen-
sions impact on cost, quality and resource needs to be determined, for each of the
system state scenarios. This means that for each possible scenario that the cogni-
tive radio would encounter, it should be determined what the impact of a certain
control know configuration would be on the cost, quality or resource dimensions.
After these four design-time phases, the run-time phase can be carried out, as will
be explained in Sect.
6.3.2
.
6.2.1 Flexibility for Energy and QoS
Control Dimensions or Knobs
(
K
i,j
): For the considered 802.11a transceiver, we
identify several control dimensions that trade-off performance for energy savings.
Our system modeling is based on an 802.11a [76] direct conversion transceiver im-
plementation with turbo coding [77] (Fig.
6.3
). Four control dimensions have a sig-
nificant impact on energy and performance for these OFDM transceivers: the mod-
ulation order (
N
Mod
), the code rate (
B
c
), the power amplifier transmit power (
P
Tx
)
and its linearity specified by the back-off (
b
).
For the energy modeling, we focus on the power amplifier control knob as PA's
generally are the most power-hungry component in the transmitter consuming up-
wards of 600 mW [78]. The major drawback for 802.11a OFDM modulation is the
large peak-to-average power ratio (PAPR) of the transmitted signal.
1
AhighPAPR
renders the implementation costly and inefficient since power-efficient PA designs
require a reduced signal dynamic range [80]. However, reducing the PA's linear
power amplification range clips the transmitted signal and increases the signal dis-
tortion.
To combat this, new PA designs allow to adapt the gain compression character-
istic independently of the output power. Adapting the PA gain compression charac-
teristic allows to translate a transmit power or linearity reduction into an effective
1
We note that digital techniques such as clipping also enable to significantly decrease the PAPR
constraint [79].
