Josef Vychodil’s research while affiliated with Brno University of Technology and other places

The spatial statistics of radio wave propagation in specific environments and scenarios, as well as being able to recognize important signal components, are prerequisites for dependable connectivity. There are several reasons why in-vehicle communication is unique, including safety considerations and vehicle-to-vehicle/infrastructure communication.The paper examines the characteristics of clustering power delay profiles to investigate in-vehicle communication. It has been demonstrated that the Saleh-Valenzuela channel model can also be adapted for in-vehicle communication, and that the signal is received in clusters with exponential decay. A measurement campaign was conducted, capturing the power delay profile inside the vehicle cabin, and the reweighted l1 minimization method was compared with the traditional k-means clustering techniques.

This paper introduces an approach to process channel sounder data acquired from Channel Impulse Response (CIR) of 60GHz and 80GHz channel sounder systems, through the integration of Long Short-Term Memory (LSTM) Neural Network (NN) and Fully Connected Neural Network (FCNN). The primary goal is to enhance and automate cluster detection within peaks from noised CIR data. The study initially compares the performance of LSTM NN and FCNN across different input sequence lengths. Notably, LSTM surpasses FCNN due to its incorporation of memory cells, which prove beneficial for handling longer series.Additionally, the paper investigates the robustness of LSTM NN through various architectural configurations. The findings suggest that robust neural networks tend to closely mimic the input function, whereas smaller neural networks are better at generalizing trends in time series data, which is desirable for anomaly detection, where function peaks are regarded as anomalies.Finally, the selected LSTM NN is compared with traditional signal filters, including Butterworth, Savitzky-Golay, Bessel/Thomson, and median filters. Visual observations indicate that the most effective methods for peak detection within channel impulse response data are either the LSTM NN or median filter, as they yield similar results.

This paper presents a comprehensive measurement campaign aimed at evaluating indoor-to-indoor radio channels in dynamic scenarios, with a particular focus on applications such as autonomous ground vehicles (AGV). These scenarios are characterized by the height of the antennas, addressing the unique challenges of near-ground communication. Our study involves long-term measurements (20 minutes of continuous recording per measurement) of the channel impulse response (CIR) in the 60 GHz and 80 GHz frequency bands, each with a bandwidth of 2.048 GHz. We investigate the variations in channel characteristics, focusing on parameters such as root mean square (RMS) delay spread and the Rician factor.

The deployment of wireless systems in millimeter wave relies on a thorough understanding of electromagnetic wave propagation under various weather conditions and scenarios. In this study, we characterize millimeter wave propagation effects from measurement data, utilizing channel impulse response analysis with a focus on root mean square delay spread and Rician K -factor. The obtained results highlight the significant influence of weather conditions and foliage on propagation, including diffuse reflection, scattering, and absorption. Particularly, we observed a notable increase in scattering from deciduous trees with leaves, in comparison with bare trees or ones covered by snow or ice. The attenuation of the signal propagated through a tree with foliage is 2.16 dB/m higher compared to a bare tree. Our validation measurements within a semi-anechoic chamber confirmed these observations and aided in quantifying the differences. These findings offer valuable insights into the dynamics of millimeter-wave signals that are important for advancing wireless communication technologies.

We present two datasets for Machine Learning (ML)-based Predictive Quality of Service (PQoS) comprising Vehicle-to-Infrastructure (V2I) and Vehicle-to-Vehicle (V2V) radio channel measurements. As V2V and V2I are both indispensable elements for providing connectivity in Intelligent Transport Systems (ITS), we argue that a combination of the two datasets enables the study of Vehicle-to-Everything (V2X) connectivity in its entire complexity. We describe in detail our methodologies for performing V2V and V2I measurement campaigns, and we provide illustrative examples on the use of the collected data. Specifically, we showcase the application of approximate Bayesian Methods using the two presented datasets to portray illustrative use cases of uncertainty-aware Quality of Service and Channel State Information forecasting. Finally, we discuss novel exploratory research direction building upon our work. The V2I and V2V datasets are available on IEEE Dataport, and the code utilized in our numerical experiments is publicly accessible via CodeOcean.