Loading Algorithms for Adaptive SS-MC-MA Systems over Wireline Channels: Comparison with DMT

In this paper, we propose to combine adaptive loading principles with the spread-spectrum multicarrier multiple access (SS-MC-MA) scheme. Such an approach has particular interests in the context of powerline communications (PLC), where the transmitter has not only to exploit robust transmission techniques, but has also to adapt the waveform to the channel response. We introduce finite-granularity loading algorithms that dynamically handle the configuration of the system under power spectral density constraints. The presented algorithms assign subcarriers, spreading codes, bits and energy to each user in order to maximize either the data rate or the noise margin at a given target symbol error rate. These algorithms can actually be viewed as a widening of the classical waterfilling approach in the case of an hybrid spread-spectrum multicarrier system. Simulation results of the new scheme are presented for different measured PLC channels and are compared with those of the classical discrete multitone modulation (DMT) approach. It is shown that the adaptive SS-MC-MA scheme performs significantly better than DMT, due to its natural energy gathering capability. Adaptive SS-MC-MA then leads to a more efficient bits and energies distribution and constitutes a simple solution to reduce the quantification loss induced by the use of finite order modulations