HOHyukjun Oh

Hyukjun Oh

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Hyukjun Oh has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Hybrid Precoding and Combining Based on Rate Balancing for Wideband mmWave Multiuser MIMO Systems

    by Jihoon Choi, Hyukjun Oh

    Considering wideband millimeter wave (mmWave) multiuser multiple-input multiple-output (MIMO) systems, we propose a new hybrid precoding and combining method for rate balancing among users. The orthogonal frequency division multiple access (OFDMA) scheme is employed for multiuser transmission, and a greedy-based algorithm is developed for subband allocation under the max-min user rate criterion. A new algorithm is derived that optimizes the radio frequency (RF) precoder and combiners under the constant-modulus constraints as well as the baseband precoders and combiners under the rate balancing power allocation criterion. The proposed algorithm iteratively updates the RF precoder and combiners using the conjugate gradient method, and baseband precoders and combiners through the singular value decomposition and numerical gradient-based power allocation among users and subcarriers. The complexity of the proposed method is compared to existing wideband hybrid processing schemes through analysis and numerical runtime measurement. The convergence of the proposed algorithm is verified via numerical simulations in mmWave MIMO-OFDMA systems. It is also shown through numerical simulations that the proposed wideband hybrid processing method outperforms the existing wideband hybrid techniques in terms of the minimum user rate, regardless of the signal-to-noise ratio, the number of antennas, and the number of users. Moreover, simulation results present that the proposed scheme is more advantageous than the conventional hybrid processing methods under channel uncertainty.


  • Analysis and classification of drone sounds from digital media

    by Sinwoo Yoo, Yeongbin Ryu, Sunghoon Shin, Hyukjun Oh

    The use of the quadcopter‐type drone is now at a mature and practical stage, and many major manufacturers are expanding their range of applications. Because of the high maneuverability and practicality of this flying object, how we ensure it is used for benign purposes in urban life has become a major issue. Most such drones are small enough to avoid many current airborne detection methods and cheap enough to be disposable. We were able to gather enough amount of multimedia sources including drone‐related sources using open online platforms, such as YouTube and other similar services. For a subset of commercial quadcopter drones, we collected sound sample data consisting of recordings and YouTube videos and built a trained simple nonlinear neural network filter to classify them. We used Mel‐frequency cepstral coefficients to reduce the amount of data as the input signal and created two‐dimensional spectrogram images to minimize the loss of temporal information in the sample, as suggested by previous research. Most of the acoustic signal energy in the drone sound existed in specific frequency bands and residual signal patterns were spread up to around 15 kHz. We designed and tested several recurrent and convolutional neural network models and one of the simple neural network models based on the LeNet architecture could classify over 10 different types of drones from acoustic data with over 97% accuracy including an additional path for the detection of general acoustic drone patterns.


  • Open Access

    Convolutional Neural Network (CNN)-Based Frame Synchronization Method

    by Eui-Rim Jeong, Eui-Soo Lee, Jingon Joung, Hyukjun Oh

    A new frame synchronization technique based on convolutional neural network (CNN) is proposed for synchronized networks. To estimate the exact packet arrival time, the receiver typically uses the correlator between the received signal and the preamble or pilot in front of the transmitted packet. The conventional frame synchronization technique searches the correlation peak within the time window. In contrast, the proposed method utilizes a CNN to find the packet arrival time. Specifically, in the proposed method, the 1D correlator output is converted into a 2D matrix by reshaping, and the resulting signal is inputted to the proposed 4-layer CNN classifier. Then, the CNN predicts the packet arrival time. To verify the frame synchronization performance, computer simulation is performed for two channel models: additive white Gaussian noise and fading channels. Simulation results show that the proposed CNN-based synchronization method outperforms the conventional correlation-based technique by 2dB.