ETRI Journal最新文献

Dynamic object detection is essential for ensuring safe and reliable autonomous driving. Recently, light detection and ranging (LiDAR)-based object detection has been introduced and shown excellent performance on various benchmarks. Although LiDAR sensors have excellent accuracy in estimating distance, they lack texture or color information and have a lower resolution than conventional cameras. In addition, performance degradation occurs when a LiDAR-based object detection model is applied to different driving environments or when sensors from different LiDAR manufacturers are utilized owing to the domain gap phenomenon. To address these issues, a sensor-fusion-based object detection and classification method is proposed. The proposed method operates in real time, making it suitable for integration into autonomous vehicles. It performs well on our custom dataset and on publicly available datasets, demonstrating its effectiveness in real-world road environments. In addition, we will make available a novel three-dimensional moving object detection dataset called ETRI 3D MOD.

Unmanned aerial vehicles (UAV) and ground vehicles (UGV) require advanced video analytics for various tasks, such as moving object detection and segmentation; this has led to increasing demands for these methods. We propose a zero-shot video object segmentation method specifically designed for UAV and UGV applications that focuses on the discovery of moving objects in challenging scenarios. This method employs a background memory model that enables training from sparse annotations along the time axis, utilizing temporal modeling of the background to detect moving objects effectively. The proposed method addresses the limitations of the existing state-of-the-art methods for detecting salient objects within images, regardless of their movements. In particular, our method achieved mean $�J$ and $�F$ values of 82.7 and 81.2 on the DAVIS'16, respectively. We also conducted extensive ablation studies that highlighted the contributions of various input compositions and combinations of datasets used for training. In future developments, we will integrate the proposed method with additional systems, such as tracking and obstacle avoidance functionalities.

The application of unmanned aerial vehicles (UAVs) has recently attracted considerable interest in various areas. A three-dimensional multiple-input multiple-output concentric two-hemisphere model is proposed to characterize the scattering environment around a vehicle in an urban UAV-to-vehicle communication scenario. Multipath components of the model consisted of line-of-sight and single-bounced components. This study focused on the key parameters that determine the scatterer distribution. A time-variant process was used to analyze the nonstationarity of the proposed model. Vital statistical properties, such as the space–time–frequency correlation function, Doppler power spectral density, level-crossing rate, average fade duration, and channel capacity, were derived and analyzed. The results indicated that with an increase in the maximum scatter radius, the time correlation and level-crossing rate decreased, the frequency correlation function had a faster downward trend, and average fade duration increased. In addition, with the increase of concentration parameter, the time correlation, space correlation, and level-crossing rate increased, average fade duration decreased, and Doppler power spectral density became flatter. The proposed model was compared with current geometry-based stochastic models (GBSMs) and showed good consistency. In addition, we verified the nonstationarity in the temporal and spatial domains of the proposed model. These conclusions can be used as references in the design of more reasonable communication systems.

Mobile robots are used in modern life; however, object recognition is still insufficient to realize robot navigation in crowded environments. Mobile robots must rapidly and accurately recognize the movements and shapes of pedestrians to navigate safely in pedestrian-rich spaces. This study proposes real-time, accurate, three-dimensional (3D) multi-pedestrian detection and tracking using a 3D light detection and ranging (LiDAR) point cloud in crowded environments. The pedestrian detection quickly segments a sparse 3D point cloud into individual pedestrians using a lightweight convolutional autoencoder and connected-component algorithm. The multi-pedestrian tracking identifies the same pedestrians considering motion and appearance cues in continuing frames. In addition, it estimates pedestrians' dynamic movements with various patterns by adaptively mixing heterogeneous motion models. We evaluate the computational speed and accuracy of each module using the KITTI dataset. We demonstrate that our integrated system, which rapidly and accurately recognizes pedestrian movement and appearance using a sparse 3D LiDAR, is applicable for robot navigation in crowded spaces.

A hardware architecture is presented to decode (N, K) polar codes based on a low-density parity-check code-like decoding method. By applying suitable pruning techniques to the dense graph of the polar code, the decoder architectures are optimized using fewer check nodes (CN) and variable nodes (VN). Pipelining is introduced in the CN and VN architectures, reducing the critical path delay. Latency is reduced further by a fully parallelized, single-stage architecture compared with the log N stages in the conventional belief propagation (BP) decoder. The designed decoder for short-to-intermediate code lengths was implemented using the Virtex-7 field-programmable gate array (FPGA). It achieved a throughput of 2.44 Gbps, which is four times and 1.4 times higher than those of the fast-simplified successive cancellation and combinational decoders, respectively. The proposed decoder for the (1024, 512) polar code yielded a negligible bit error rate of 10⁻⁴ at 2.7 E_b/N_o (dB). It converged faster than the BP decoding scheme on a dense parity-check matrix. Moreover, the proposed decoder is also implemented using the Xilinx ultra-scale FPGA and verified with the fifth generation new radio physical downlink control channel specification. The superior error-correcting performance and better hardware efficiency makes our decoder a suitable alternative to the successive cancellation list decoders used in 5G wireless communication.

Even though generalized frequency division multiplexing is an alternative waveform method expected to replace the orthogonal frequency division multiplexing in the future, its implementation must alleviate channel effects. Least-squares (LS), a low-complexity channel estimation technique, could be improved by using the discrete Fourier transform (DFT) without increasing complexity. Unlike the usage of the LS method, the DFT-based method requires the receiver to know the channel impulse response (CIR) length, which is unknown. This study introduces a simple, yet effective, CIR length estimator by utilizing LS estimation. As the cyclic prefix (CP) length is commonly set to be longer than the CIR length, it is possible to search through the first samples if CP is larger than a threshold set using the remaining samples. An adaptive scale is also designed to lower the error probability of the estimation, and a simple signal-to-interference-noise ratio estimation is also proposed by utilizing a sparse preamble to support the use of the scale. A software simulation is used to show the ability of the proposed system to estimate the CIR length. Due to shorter CIR length of rural area, the performance is slightly poorer compared to urban environment. Nevertheless, satisfactory performance is shown for both environments.

This paper demonstrates improved throughput and energy efficiency of wireless communications by exploiting nonorthogonal multiple access (NOMA), multiple input–multiple output (MIMO), and radio frequency energy harvesting (EH) technologies. To assess the performance of NOMA MIMO communications with EH (MMe), we consider the nonlinear characteristics of EH devices and propose explicit expressions for throughput and outage probability. Based on our results, the system performance is significantly mitigated by EH nonlinearity and is considerably improved by increasing the number of antennas. Additionally, by appropriately adjusting the system parameters, our NOMA MMe innovation can avert complete outages while optimizing system performance. Moreover, the results demonstrate the superiority of the NOMA MMe over its orthogonal multiple access MMe counterparts.

A compact triple-band porous electromagnetic bandgap structure-loaded coplanar-waveguide-fed wearable antenna is introduced for applications of wireless body area networks. The porous structure is aimed to create a stopband or bandgap in the electromagnetic spectrum and increase breathability. The holes in the bottom electromagnetic bandgap surface increase the inductance, which in turn increases the bandwidth. The final design resonates at three bands with impedance bandwidths of 264 MHz, 100 MHz, and 153 MHz and maximum gains of 2.18 dBi, 6.75 dBi, and 9.50 dBi at 2.45 GHz, 3.5 GHz, and 5.5 GHz, respectively. In addition, measurements indicate that the proposed design can be deformed up to certain curvature and withstand human tissue loading. Moreover, the specific absorption rate remains within safe levels for humans. Therefore, the proposed antenna can suitably operate in the industrial, scientific, and medical, Bluetooth, Wi-Fi, and WiMAX bands for potential application to wireless body area networks.

Advancements in remote medical technologies and smart devices have led to expectations of contactless rehabilitation. Conventionally, rehabilitation requires clinicians to perform routine muscle function assessments with patients. However, assessment results are difficult to cross-reference owing to the lack of a gold standard. Thus, the application of remote smart rehabilitation systems is significantly hindered. This study analyzes the factors affecting the real-time evaluation of muscle function based on biometric sensor data so that we can provide a basis for a remote system. We acquired real clinical stroke patient data to identify the meaningful features associated with normal and abnormal musculature. We provide a system based on these emerging features that assesses muscle functionality in real time via streamed biometric signal data. A system view based on the amount of data, data processing speed, and feature proportions is provided to support the production of a rudimentary remote smart rehabilitation system.

We propose a network traffic prediction model based on linear and nonlinear model combination. Network traffic is modeled by an autoregressive moving average model, and the error between the measured and predicted network traffic values is obtained. Then, an echo state network is used to fit the prediction error with nonlinear components. In addition, an improved slime mold algorithm is proposed for reservoir parameter optimization of the echo state network, further improving the regression performance. The predictions of the linear (autoregressive moving average) and nonlinear (echo state network) models are added to obtain the final prediction. Compared with other prediction models, test results on two network traffic datasets from mobile and fixed networks show that the proposed prediction model has a smaller error and difference measures. In addition, the coefficient of determination and index of agreement is close to 1, indicating a better data fitting performance. Although the proposed prediction model has a slight increase in time complexity for training and prediction compared with some models, it shows practical applicability.