Broadband PLC channel estimation and equalization techniques

Abstract

Over the last years, high-speed broadband power line communications (BB-PLC) over low voltage networks have become an option worth of consideration for residential data transmission. This technology have specific standards that take into account the modulation technique, among other things. Multi-carrier systems, like orthogonal frequency-division multiplexing (OFDM), or alternatively the use of single-carrier are mainly considered. Both modulation techniques require the use of channel equalization techniques in order to mitigate the effect of the physical medium on the transmitted signal, which behaves like a time-variant multipath channel that severely distorting the transmitted signal. The equalization can be performed in the time or frequency domain, with time domain equalization techniques often requiring more processing load than frequency domain techniques. To perform frequency domain equalization (FDE), a cyclic prefix is added to the transmitted signal to transform the effect ofthe channel into a periodic signal, allowing to process the data packet with a Discrete Fourier Transform (DFT), minimizing the calculation errors as long as the prefix is longer than the channel impulse response. A crucial part of the equalization process is the correct identification of the channel response, as it is necessary in order to obtain a correct equalization term. There are several techniques in the literature that allow to obtain the channel response dynamically, which is required because the channel is time-variant. One of these techniques is the use of Complementary Sets of Sequences (CSS) as pilot symbols. These sequences have the property that the sum of autocorrelation functions of the sequences produces a Kronecker delta, which allows to obtain the discrete impulse response of the channel by transmitting them over the medium. This chapter will show frequency domain equalization techniques based on CSS for single-carrier and OFDM. These techniques will be compared to well known techniques like least squares (LS) and Linear Minimum Mean-Square Error (LMMSE) for OFDM, and Zero-Forcing (ZF) and Minimum Mean-Square Error (MMSE) for single carrier. The comparisons will be done considering the Bit Error Rate (BER) with respect to Signal-to-Noise ratio (SNR) of transmissions over a realistic simulation of PLC channels. The channel simulation is composed by a time-varying channel response model and Middleton Class-A noise, which is considered as the model closer to the real PLC noise. The shown methods are also compared over a real data transmission, with the number of received errors as comparison metric. Finally a comparison between the OFDM and single-carrier techniques is shown.