Techniques for broadband power line communications: impulsive noise mitigation and adaptive modulation

The development of power line communication systems for broadband multimedia applications requires a comprehensive knowledge of the channel characteristics and the main peculiarities that may influence the communication over this channel. PLC has the potential to become the preferred connectivity solution to homes and offices. Additionally, indoor power line networks can serve as local area networks offering high-speed data, audio, video and multimedia applications. The PLC technology eliminates the need for new wires by using an already-existing infrastructure that is much more pervasive than any other wired system. Power line networks, however, present a hostile channel for communication signals. Noise, multipath, selective fading and attenuation are well-known peculiarities of power line grids and. Particularly, random impulsive noise characterized with short durations and very high amplitudes is identified as one of the major impairments that degrade the performance of PLC systems. Orthogonal frequency division multiplexing (OFDM) is the technique of choice for broadband PLC systems. OFDM minimizes the effects of multipath and provides high robustness against selective fading. It is also powerful in impulsive noise environments and performs better than single-carrier modulation methods. If an OFDM symbol is affected by impulsive noise, the effect is spread over multiple subcarriers due to the discrete Fourier transform at the receiver. In order to achieve reliable outcomes, suitable channel and noise models must be used in the investigations. In this thesis, the power line channel transfer function is modelled using a multipath model that was proposed by Zimmermann and Dostert [1], [2]. This model describes the signal propagation scenario and attenuation effects in power line networks. To represent the actual noise scenario in power networks, the noise is classified into two main classes: background noise and impulsive noise. To reduce the effect of impulsive noise, conventional time domain nonlinearities are examined in this thesis under PLC environments. An adaptive-threshold selection method based on minimum bit-error rate (BER) is proposed. At the cost of additional complexity, the effect of impulsive noise is further mitigated using a novel joint time-domain/frequency-domain suppression technique. Since channel coding is essential for most telecommunication systems, we examine convolutional codes combined with interleaving in a PLC channel impaired with AWGN and impulsive noise. The results show substantial performance gains especially in heavily-disturbed environments, where signal-to-noise ratio (SNR) gains of more than 15 dB can be achieved with a code rate of 1/3. Bit-interleaved convolutionally-coded OFDM completely eliminates the effect of impulsive noise in weakly-disturbed noise environments, while a negligible effect may remain in medium-disturbed environments. A new power-loading algorithm that minimizes the transmission power for target BER and data rate constraints is introduced in later chapters of the thesis. Results indicate that the algorithm achieves performance gains of more than 4 dB SNR over conventional OFDM systems. Furthermore, a novel minimum-complexity bit-loading algorithm that maximizes the data rate given BER and power level constraints is proposed in chapter 6. Results show that this bit-loading algorithm achieves almost identical performance as the incremental algorithm but with much lower complexity