• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.1230-1233
• /
• 2023

최근, 65세 이상의 1인 가구가 급증함에 따라 노인을 대상으로 한 다양한 연구 및 서비스가 활발히 이루어지고 있다. 이에 본 논문에서는 시계열 센서 데이터를 이용하여 CNN-LSTM 기반의 노인 활동 인식 시스템을 제안한다. 수집된 데이터는 3축 가속도 센서가 내장된 2개의 디바이스를 등과 허벅지에 부착하였다. 수집 주기는 50hz로 진행되었으며, 각 행동은 2초를 기준으로 산정하였다. 학습데이터의 입력값으로 사용하기 위해, 슬라이딩 윈도우를 50%로 적용하여 시퀀스를 구성하였다. 모델은 특징을 반영하기 위한 CNN(Convolutional Neural Networks)과 시계열적 특성을 반영하기 위한 LSTM(Long-Short Term Memory)을 하이브리드한 1차원 형태의 CNN-LSTM 모델을 사용한다. 행동은 4가지로 분류하였으며, 97%의 정확도를 나타내고 있다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.24 no.8A
• /
• pp.1238-1244
• /
• 1999

In this paper, a new design of a triple-erroe-correcting (TEC) Reed-Solomon decoder is presented based on direct decoding method which is more efficient for the case of relatively small error correction capability. The proposed decoder requires only 9 GF(2m) multipliers in obtaining the error-locator polynomial and the error-evaluator polynomial, whereas other decoders needs 24 multipliers. Thus, the attractive feature of this decoder is its remarkable simplicity from the point of view of implementation. Futhermore, the proposed TEC Reed-Solomon decoder has very simple control circuit and short decoding delay. Therefore this decoder can be implemented by simple hardware and also save buffer memory which stores received sequence.

• Journal of Broadcast Engineering
• /
• v.25 no.4
• /
• pp.576-586
• /
• 2020

Analyzing baseball games using sensors such as radars or riders is expensive. In this paper, we propose an algorithm to detect pitch shots and hit shots using baseball video and to generate ball trajectories within hit shots using camera movement. After the pitch shot and the hit shot detection using object detection and optical flow, we generate the transformation relationship between frames and ball locations in the frame, and calculates the ball trajectory. The performance of the proposed method is evaluated for three KBO baseball video sequences, and the detection accuracy and detection rate of pitch shot and hit shot were within 89-95 [%], and the average error for shot range was 13.6[m], The direction error was 7.5° and foul classification accuracy was 98.6%.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.10C
• /
• pp.992-998
• /
• 2005

In this paper, An Ultra Fast Acquisition Algorithm of low transmission rate ultra-wideband(UWB) systems for car positioning and traffic light controling is proposed. Since the acquisition algorithms for CDMA system are not fast enough to access the low transmission rate UWB systems, the new ultra fast acquisition scheme which can be implemented with low cost and simplified circuit is required. The proposed algorithm adopted the Recurrent Sequential Estimation scheme and trinomial M-sequence. Therefore, The proposed scheme can reduce the average acquisition time in $1\~3{\mu}sec$ with simple circuit, even for the UWB systems which use long pseudo-noise(PN) sequence and transmit low power below the FCC EIRP emission limits. The simulation results for the average acquisition time of the proposed scheme are compared with the ones of the existing acquisition schemes.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.9
• /
• pp.2080-2089
• /
• 1997

Recent evolution in compute technology makesthe multimedia processing widely availiable. Conventional storage systems do not meet the requirements of multimedia data. Several approaches were suggested to improve disk storing methods for them. Bocheck proposed a disk partitioning technique for multiple steams assuming that all steams have same retrieval intervals with the same amount data for each access. While Bocheck's one provides a good method for same period, it does not consider the case of different periods of continous media streams. This paper proposes a new partitioning technique in which a fixed number of blocks are assigned for stresms with different retrieval periodicity. The analysis shows this problem is the same as the one scheduling the steams into a given sequence. The simulation was done to compare the proposed m-sequence merge method with the conventional Scan-EDF and Partitioning methods.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.4C
• /
• pp.191-196
• /
• 2011

In this paper, we propose a design method of Reed-Solomon (23, 17) decoder for UWB using direct decoding method. The direct decoding algorithm is more efficient for the case of relatively small error correction capability. The proposed decoder requires only 9 $GF(2^m)$ multipliers in obtaining the error-locator polynomial and the error-evaluator polynomial, whereas other decoders need about 20 multipliers. Thus, the attractive feature of this decoder is its remarkable simplicity from the point of view of hardware implementation. Futhermore, the proposed decoder has very simple control circuit and short decoding delay. Therefore this decoder can be implemented by simple hardware and also save buffer memory which stores received sequence.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.3 no.1
• /
• pp.129-139
• /
• 1999

DECT(Digital Enhanced Cordless Telecommunications) is the communication protocol and the pico-cellular network that provides a PABX with mobile services on the local area within its cell radius of about 100 to 200m. Compared with GSM, DECT provides low-cost communication services based on the PSTN and it is an ideal supplementary method for the type of in-building mobile communication. In order to implement the data link communication layer of the DECT system, in this paper we analyzed and designed the structures and functions of it that provided reliable communication and real-time services to the upper network layer by performing the functions of link controls, error controls, sequence controls, and flow controls.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.5A
• /
• pp.479-486
• /
• 2008

In this paper, an efficient training symbol design with m-sequence is proposed for the MIMO-OFDM based next generation wireless transmission system which supports gigabits per second data rate. In the traditional blute force method, the preamble design is based on the case by case comparison with the system requirements. This paper discusses a training symbol design methodology for the MIMO-OFDM system based on the m-sequence which has been widely used in the spread spectrum communication areas due to its good correlation characteristics. Also the step-by-step design and performance verification method within the limited search space is discussed. The proposed method targets the design of the training symbol which satisfies system requirements for the packet based MIMO-OFDM wireless communication system including automatic gain control(AGC), timing synchronization, frequency and sampling offset estimation, and MIMO channel estimation.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.6
• /
• pp.57-65
• /
• 2018

Conventional RF signal-based indoor localization techniques such as BLE or Wi-Fi based fingerprinting method show considerable localization errors even in small-scale indoor environments due to unstable received signal strength(RSS) of RF signals. Therefore, it is difficult to apply the existing RF-based fingerprinting techniques to large-scale indoor environments such as airports and department stores. In this paper, instead of RF signal we use the geomagnetic sensor signal for indoor localization, whose signal strength is more stable than RF RSS. Although similar geomagnetic field values exist in indoor space, an object movement would experience a unique sequence of the geomagnetic field signals as the movement continues. We use a deep neural network model called the recurrent neural network (RNN), which is effective in recognizing time-varying sequences of sensor data, to track the user's location and movement path. To evaluate the performance of the proposed geomagnetic field based indoor positioning system (IPS), we constructed a magnetic field map for a campus testbed of about $94m{\times}26$ dimension and trained RNN using various potential movement paths and their location data extracted from the magnetic field map. By adjusting various hyperparameters, we could achieve an average localization error of 1.20 meters in the testbed.

• Economic and Environmental Geology
• /
• v.38 no.6 s.175
• /
• pp.633-642
• /
• 2005

We present the sedimentary sequence and distribution pattern of the late Holocene muddy deposits in the northern East China Sea shelf using the high-resolution 'Chirp' profiles. The seismic sedimentary sequence overlying acoustic basement (basal reflector-B) can be divided into two depositional units (Unit 1 and 2) bounded by erosional bounding surface (mid reflector-M). The lower Unit 1 above basal reflector-H is characterized by the acoustically parallel to subparallel reflections and channel-fill facies. The upper Unit 2, up to 7 m in thickness, shows seismically semi-transparent seismic facies and lenticular body form. On the base of sequence stratigraphic concept, these two sediment units have developed during transgression and highstand period, respectively, since the last sea-level lowstand. The transgressive systems tract (Unit 1) lie directly on the sequence boundary (reflector B) that have farmed during the last glacial maximum. The transgressive systems tract in this study consists mostly of complex of delta, fluvial, and tidal deposits within the incised valley estuary system. The maximum flooding surface (reflector M) corresponding to the top surface of transgressive systems tract is obviously characterized by erosional depression. The highstand systems tract (Unit 2) above maximum flooding surface is made up of the mud patch filled with the erosional depression. The high-stand mud deposits showing a circle shape just like a typhoon symbol locates about 140 km off the south of Cheju Island with water depth of $60\~90m$. Coverage area and total sediment volume of the mud deposits are about $3,200km^2$ and $10.7\times10^9\;m^3$, respectively. The origin of the mud patch is interpreted as a result of accumulating suspended sediments derived from the paleo-Yellow and/or Yangtze Rivers. The circular distribution pattern of the mud patch appears to be largely controlled by the presence of cyclonic eddy in the northern East China Sea.