Invoking Deep Learning for Joint Estimation of Indoor LiFi User Position and Orientation

Light-fidelity (LiFi) is a fully-networked bidirectional optical wireless communication (OWC) technology that considered as promising solution for high-speed indoor connectivity. In this paper, the joint estimation of user 3D position and equipment (UE) orientation in LiFi systems with unknown emission power investigated. Existing solutions problem assume either ideal system settings or perfect knowledge UE states, rendering them unsuitable realistic systems. addition, these consider non-line-of-sight (NLOS) links channel gain source deterioration performance instead harnessing components improving performance. This mainly due to lack appropriate techniques can extract information hidden components. against above limitations, assumed be connected at least one access point (AP), i.e., active link. Fingerprinting employed an technique received signal-to-noise ratio (SNR) used metric, where both line-of-sight (LOS) NLOS are considered. Motivated by success deep learning solving several complex prediction problems, we employ two artificial neural network (ANN) models, based on multilayer perceptron (MLP) second convolutional (CNN), map efficiently instantaneous SNR orientation. Through numerous examples, investigate proposed schemes terms average error, precision, computational time, bit error rate. We also compare k-nearest neighbours (KNN) scheme, which widely localization problems. It demonstrated achieve significant gains superior KNN scheme.