Study of MIMO techniques for Power Line Communications

• Autores: Julio Alberto Corchado López
• Directores de la Tesis: Luis Díez del Río (dir. tes.), José Antonio Cortés Arrabal (codir. tes.)
• Lectura: En la Universidad de Málaga ( España ) en 2024
• Idioma: inglés
• Tribunal Calificador de la Tesis: Fernando Cruz Roldán (presid.), María Carmen Aguayo Torres (secret.), Francisco Javier López Martínez (voc.)
• Programa de doctorado: Programa de Doctorado en Ingeniería de Telecomunicación por la Universidad de Málaga
• Enlaces
  • Tesis en acceso abierto en: RIUMA
• Resumen

  • Power line communication (PLC) consists in the exchange of information over electrical cables. PLC takes advantage of the ubiquity of already deployed power delivery networks and provides access to telecommunication services without any further infrastructure installation. Furthermore, the propagation of PLC signals is insensitive to wall/floor thickness, therefore they can propagate in multi-storey premises much better than wireless.

    PLC applications are generally categorized as either outdoor or indoor. Outdoor applications usually take place over the distribution power network and the signal travel distance may be of up to a few kilometers. A well-known application of outdoor PLC systems is smart metering. On the other hand, indoor applications are usually focused on, but not restricted to, providing high data rate connectivity inside buildings. In both cases, PLC can complement wireless communication systems to improve coverage. Moreover, PLC systems can also be classified into narrowband (NB) and broadband (BB) systems. The work presented herein focuses on the latter.

    Traditionally, PLC systems used only two conductors, thus resulting in a singleinput single-output (SISO) system. However, most indoor power lines are composed of three conductors, and this third conductor can be exploited to enable a multiple-input multiple-output (MIMO) system over this channel. Exploitation of MIMO capabilities enhance the achievable performance over a given medium. MIMO features have been extensively studied for wireless environments but MIMO in wired scenarios show some distinct aspects with respect to their wireless counterparts. In particular, the channels that make up the MIMO PLC link show a closer relation (higher spatial correlation) than in their wireless counterpart. This higher spatial correlation can be observed in both, the channel response and noise of the MIMO PLC channel. Regarding the channel response, a higher spatial correlation entails a lower MIMO performance gain, whereas highly correlated noise can be exploited to achieve higher MIMO performance gains. However, nowadays the channel spatial correlation dependence with frequency (both in the channel response and noise) is unknown, as it is unknown the relationship of the channel response spatial correlation to the physical features of the network, such as the type of wiring, the type of deployment, etc. Furthermore, existing MIMO channel response models do not reflect an accurate representation of the spatial correlation observed in measurements. MIMO PLC models and transmission strategies need to take into account these particularities to take full advantage of MIMO capabilities.

    The first contribution of this work is a multiconductor transmission line (MTL)-based MIMO PLC model which captures spatial correlation in a way that other MTL-based models cannot. Three modifications on the channel model are proposed to achieve this: modified loads, branches and cabling. The proposed model leads to MIMO channels similar to the measured ones, both in terms of the characteristics of their individual SISO channels and of the correlation between them.

    The second contribution shows, by means of measurements, that spatial correlation in PLC channel responses does not display a significant dependence with respect to frequency, just like in the typical wireless MIMO scenario, and that there exist alternative injection modes to the usual differential one, like the pseudodifferential injection, which yield lower spatially correlated MIMO channels.

    Lastly, the third contribution provides a characterization of the noise correlation in MIMO PLC in the 2-108 MHz band, showing that noise in the frequency modulation (FM) band displays very high spatial correlation. It has been proven that the exploitation of the aforesaid correlation, by means of a linear precoding system with a whitening transformation, makes PLC feasible in the FM band with an injected power spectral density (PSD) as low as -100 dBm/Hz (enough to avoid interference with FM signal reception systems).