A new loading algorithm for discrete multitone transmission

Abstract

A new loading algorithm for discrete multitone transmission is proposed. Thereby rate is not distributed according to channel capacity, but rate and transmit power are assigned to maximize the signal-to-noise ratio in each carrier. Because closed form expressions can be derived the algorithm is of very low complexity, even lower than the loading algorithm recently proposed by Chow et al. [ 3 ] . Nevertheless achievable performance is higher or at least the same. Results for a typical high rate transmission over twisted pair lines are presented.

Adaptive BICM-OFDM Systems

“…In order to carry out the optimisation, the rates have to be formulated as explicit functions of the powers, such that r k = f (p k ), which in case of coded systems with finite symbol alphabets is a non-trivial problem. If channel coding is neglected, analytical BER/FER expressions, e.g., for M -QAM modulations, can be used [1][2][3]. Due to the difficulties in finding analytical expressions for the FER of coded systems, however, the simulated performance of different code and modulation combinations is often utilised to quantify the performance [8,10].…”

“…Early works focused on uncoded BER optimisation [1][2][3], but nowadays the consideration of channel coding into bit and power loading schemes has become a major aspect. Several contributions to this field have been made, especially considering the influence of Bit Interleaved Coded Modulation (BICM) on such systems.…”

Turbo-coding and Bit-loading Algorithms for a HomePlug-like DMT PLC System

“…It can be observed that the algorithm tends to saturate the best channels, i.e., the ones with better channel conditions. Moreover, since the number of bits per symbol cannot be greater than 2, overall bit-rate tends to an asymptotic value, which in our case was about 30 Mbls and 10 Mbls for the coherent and differential case, respectively.…”

“…Fischer's algorithm [10] is the most recent of three. It aims at minimizing the probability of error in each subchannel: this algorithm relies on a set of iterative equations, which lead to a flat-energy distribution, and show a slight improvement in SNR over Chow solution.…”

“…As it is known, the optimal Bit-loading scheme is based on the water-pouring distribution: it can be shown that water-filling algorithm converges to Shannon channel capacity as the bandwidth of the subchannels gets smaller [6]. In practical applications, however, the optimal solution cannot be determined, because it assumes infinite granularity in constellation sizes and in sub band division, which are not realizable: several sub-optimal bit-loading algorithms have been recently investigated in literature [4,[7][8][9][10].…”