no code implementations • 24 Jul 2023 • Sung Joon Maeng, Hyeokjun Kwon, Ozgur Ozdemir, İsmail Güvenç
By analyzing reference signal received power (RSRP) using path loss models that account for antenna patterns, we observed that our measurement results, obtained at different UAV heights, aligned well with the two-ray path loss model when incorporating the measured antenna pattern.
no code implementations • 28 May 2023 • Hyeokjun Kwon, Sung Joon Maeng, Ismail Guvenc
In this paper, the efficacy of the proposed localization approaches is examined based on these two main localization requirements through computer simulations.
no code implementations • 17 Jan 2023 • Hyeokjun Kwon, Ismail Guvenc
The use of unmanned aerial vehicles (UAVs) for signal source search and localization (SSSL) can have significant advantages when compared to terrestrial-based approaches, due to the ease of capturing RF signals at higher altitudes and the autonomous 3D navigation capabilities of UAVs.
1 code implementation • NeurIPS 2021 • Yonghyeon LEE, Hyeokjun Kwon, Frank Park
Unlike existing graph-based methods that attempt to encode the training data to some prescribed latent space distribution -- one consequence being that only the encoder is the object of the regularization -- NRAE merges local connectivity information contained in the neighborhood graphs with local quadratic approximations of the decoder function to formulate a new neighborhood reconstruction loss.