Visible Light Communication is considered one of the most promising solutions to provide wireless connectivity in indoor scenarios. Furthermore, the simultaneous support of illumination and communication services makes such technology perfectly suited to the paradigm of smart environments. Data
communication is obtained by modulating the optical signal emitted by a Light Emitting Diode. Taking advantage of spatial diversity, multiple sources can be also employed to improve the transmission rate.
However the arising co-channel interference, if not properly tackled, may significantly impact on the system reliability. In this direction, in this project entitled "Optical Wireless Modulation Constellation Optimization" (OWMCO) , we aim at investigating a Multiple-Input Single-Output scenario where sources
emit spatially multiplexed, Pulse Amplitude Modulated signals. Having the receiver as equipped with a single photodetector makes signals separation and decoding very challenging, especially when channels are correlated. To this aim, the goal of this project is to propose novel modulation precoding schemes, relying on adaptive power allocation, that consider co-channel interference as a parameter to be minimized by possibly maximizing the minimum Euclidean distance between received symbols. Such techniques are expected to lead to a joint optimal constellation design and channel equalization lowering the symbol error probability while meeting the illumination constraints as well.
Specifically, modulation precoding may be performed imposing the received constellation symbols to be equally spaced. In this regard, we turned the interference cancellation into a max-min problem, with the solution requesting the maximization of the minimum Euclidean distance between symbols. The proposed mechanism first performs the LEDs power allocation by taking into account only the the illumination constraints imposed by the system (the number of potentially valid solutions is limited).
Thereafter, an iterative search among the obtained LEDs power sets is done to find a power allocation meeting the symbols equidistance constraint as well. The main advantage of the equi-spacing symbol with modulation precoding may be that the power allocation providing received symbols equidistance is searched among a limited set of possibilities, with the computational effort being significantly reduced with respect to an exhaustive search approach.
However, the constraint on symbols equidistance it is expected not to allow the full exploitation of the constellation space. In this direction, we aim at proposing an improved modulation precoding scheme, channel state information aided, aiming to achieve an optimal spacing between received symbols, no more constrained to equidistance, so that co-channel interference is minimized.
The removal of the constraint on equidistance allows the received constellation space to be exploited without any limitation.
By doing so, the number of potentially valid LEDs power sets increases, so the mechanism for equi-spaced symbols may become costly. So, we aim also at implementing a novel recursive algorithm that follows a different approach. In fact, first power allocation on LEDs is performed in order to maximize the distribution of symbols in the received constellation space. Finally, whenever needed, some power compensation may be needed in order to meet the lighting constraints as well. The goal will be also to propose a convergent solution using the minimum number of iterations so as to limit the computational effort as further detailed. Significant improvements in terms of communication performance are expected it that direction.