Implementation of channel equalization and intrusion detection in VLC systems

Maintaining the efficiency of Optical Wireless Communication (OWC) systems with channel equalization is a curial problem. In this paper, different adaptive channel estimation algorithms are studied to allow optical wireless channel equalization to deal with the physical impairments in Free Space (FS) channel. The adaptive channel estimation algorithms considered are standard Least Mean Square (LMS) and LMS with Activity Detection Guidance (ADG) and Tap Decoupling (TD). Random inputs are used in the channel estimation process. These inputs are of white and color natures. Simulation results prove that the LMS algorithm with ADG is fair enough for the channel estimation problem for white inputs. On the other hand, for colored inputs, the TD is required for correct estimation of channel taps. The asymptotic error performance of the channel estimates is directly proportional to the product of step size, number of estimated taps and noise variance. So, evaluation of the proposed scheme is introduced with unit noise variance, 0.07 and 0.01 noise variance with fixed correlation factor. Effect of changing the correlation factor and number of iterations on the estimator performance to remove channel distortion is studied, when the correlation factor is equal to 0.1, 0.5, 0.9 with fixed noise variance. Thus, we could improve and get the best asymptotic performance and convergence rate of the proposed channel estimation. In addition, he work in this paper is extended to the idea of intrusion detection that can be implemented in OWC communication systems. For this objective, we present a Cumulative Histogram (CH)-based Feature Selection (FS) model to enhance any classification process applied on the traffic transmitted over the OWC channel to alert about intrusion attempts.