Fast Fading Influence on the Deep Learning-Based LOS and NLOS Identification in Wireless Body Area Networks

User Orientation Detection in Relation to Antenna Geometry in Ultra-Wideband Wireless Body Area Networks Using Deep Learning

“…It should be emphasized that this research does not fully cover the full scope of the problem of user orientation detection, particularly with the UWB technique. In the future, one could carry out research for different dynamic scenarios, where the influence of fast fading is anticipated to influence the effectiveness of the proposed approach [3]. The range of potential study issues includes taking into account people with different body builds [4], performing channel impulse response sampling in pre-processing before delivering the data to the input of the neural network [39], placing different additional obstacles and objects between [21] and around the user and also studying the impact of other types of neural networks and hyperparameter values on user orientation detection in relation to the antenna geometry [46].…”

“…Taking into consideration the previously mentioned results and conclusions, as well as the advantages of deep learning approaches in terms of direct visibility condition detection over other methods [3,4], it was decided that deep learning methods would be analyzed in the subsequent research.…”

“…This research focuses on the analysis of the estimated channel impulse response (CIR) in order to determine the user's orientation in relation to the antenna geometry using a deep learning approach. The presented approach should be treated as an extension of the research described in [3,4]. Its novelty, in comparison to the current state of the art, mainly involves determining the direct angle of the user's position and increasing the effectiveness of determining the user orientation using convolution networks and CIR estimation.…”