Indoor Localization based on Multiple Neural Networks

손인수,Sohn, Insoo Institute of Control 2015 제어·로봇·시스템학회 논문지 Vol.21 No.4

Indoor localization is becoming one of the most important technologies for smart mobile applications with different requirements from conventional outdoor location estimation algorithms. Fingerprinting location estimation techniques based on neural networks have gained increasing attention from academia due to their good generalization properties. In this paper, we propose a novel location estimation algorithm based on an ensemble of multiple neural networks. The neural network ensemble has drawn much attention in various areas where one neural network fails to resolve and classify the given data due to its' inaccuracy, incompleteness, and ambiguity. To the best of our knowledge, this work is the first to enhance the location estimation accuracy in indoor wireless environments based on a neural network ensemble using fingerprinting training data. To evaluate the effectiveness of our proposed location estimation method, we conduct the numerical experiments using the TGn channel model that was developed by the 802.11n task group for evaluating high capacity WLAN technologies in indoor environments with multiple transmit and multiple receive antennas. The numerical results show that the proposed method based on the NNE technique outperforms the conventional methods and achieves very accurate estimation results even in environments with a low number of APs.

Iterative Decoding for LDPC Coded MIMO-OFDM Systems with SFBC Encoding

손인수,Sohn Insoo The Korean Institute of Communications and Informa 2005 韓國通信學會論文誌 Vol.30 No.5a

A multiple input multiple output orthogonal frequency division multiplexing (MIMO-OFDM) system using low-density parity-check (LDPC) code and iterative decoding is presented. The iterative decoding is performed by combining the zero-forcing technique and LDPC decoding through the use of the 'turbo principle.' The proposed system is shown to be effective with high order modulation and outperforms the space frequency block code (SFBC) method with iterative decoding. 본 논문에서는 주파수공간블록부호 (Space Frequency Block Code)를 적용한 다중송수신(Multiple Input Multiple Output)-OFDM(Orthogonal Frequency Division Multiplex) 시스템을 위한 반복복호 기법을 제시한다. 제시된 반복복호 기법은 zero-forcing 알고리즘과 LDPD 부호화 알고리즘을 터보 기법을 바탕으로 상호보완을 할 수 있도록 설계되었다. 시뮬레이션 결과에 의하면 고차원 변조방식을 적용한 MIMO-OFDM 페이딩 채널에서 기존의 시스템과 비교하여 향상된 성능을 확인하였다.

Indoor Localization based on Multiple Neural Networks

Insoo Sohn(손인수) 제어로봇시스템학회 2015 제어·로봇·시스템학회 논문지 Vol.21 No.4

Indoor localization is becoming one of the most important technologies for smart mobile applications with different requirements from conventional outdoor location estimation algorithms. Fingerprinting location estimation techniques based on neural networks have gained increasing attention from academia due to their good generalization properties. In this paper, we propose a novel location estimation algorithm based on an ensemble of multiple neural networks. The neural network ensemble has drawn much attention in various areas where one neural network fails to resolve and classify the given data due to its" inaccuracy, incompleteness, and ambiguity. To the best of our knowledge, this work is the first to enhance the location estimation accuracy in indoor wireless environments based on a neural network ensemble using fingerprinting training data. To evaluate the effectiveness of our proposed location estimation method, we conduct the numerical experiments using the TGn channel model that was developed by the 802.11n task group for evaluating high capacity WLAN technologies in indoor environments with multiple transmit and multiple receive antennas. The numerical results show that the proposed method based on the NNE technique outperforms the conventional methods and achieves very accurate estimation results even in environments with a low number of APs.

디즈니 애니메이션의 반복적 색채사용에 의한 캐릭터 강화전략에 관한 연구

손인수(Insoo Sohn),이유연(Youyeon Lee) 한국디자인학회 2002 디자인학연구 Vol.- No.46

메타버스 플랫폼을 활용한 다문화 융합교육 방안 연구

손인수(Insoo Sohn),서은숙(Eun Sook Seo) 한국통신학회 2024 한국통신학회 학술대회논문집 Vol.2024 No.1

On Modeling and Simulation of AI-based IoT

Mitra Pooyandeh,Insoo Sohn(손인수) 한국통신학회 2021 한국통신학회 학술대회논문집 Vol.2021 No.6

Nowadays, the Internet of Things has become one of the essential technologies for connecting devices in the telecommunications, electricity, medical, automotive, and other industries. The Internet of Things includes an infinite set of connections and intelligent endpoints such as sensors, actuators, and so on. IoT devices, which include sensor networks with an unlimited number of sensors, face many problems, including power limitations and hardware limitations. Therefore, simulation is the best solution for changing and extending protocols in sensor networks. Furthermore, AI is becoming an important tool to solve problem related to IoT and hardware and power constraints. In this article, we review existing simulation and modeling tools and discuss including AI in the simulation.

Study of Datasets in AI Based Medical Informatics

Samaneh Shamshiri,Insoo Sohn(손인수) 한국통신학회 2021 한국통신학회 학술대회논문집 Vol.2021 No.6

In recent years, Artificial intelligence (AI) has played a significant role in health informatics and medical applications to promote early detections, disease diagnosis, and referral managements. In terms of classification, labeling, training process, dataset size, and algorithm validation of AI, uncourtly, data is the first step to developing any treatment or detecting tools. This paper reviews some dataset analytic role and data driven models in medical informatics and investigated the two first datasets from CT scan images and chest X-ray images for detecting COVID-19 by AI models.

HAPS 기반의 HSDPA 시스템 성능 분석

김남겸(Namkyeom Kim),손인수(Insoo Sohn),이진구(Jingu Lee) 대한전자공학회 2008 電子工學會論文誌-TC (Telecommunications) Vol.45 No.6

현대에는 언제 어디서나 자신이 원하는 데이터를 빠르게 접할 수 있는 서비스가 요구되고 있고 이것을 반영하듯이 많은 시스템들이 개발되었다. 특히 HSDPA (High Speed DATA Packet Access)의 경우 최대 14.4Mbps의 데이터 수율을 가져 3세대 이동통신으로 각광을 받고 있다. 하지만 HSDPA의 경우도 지상망 시스템인 관계로 LOS (Line of Sight)가 보장되지 못해 이것은 다중 페이딩에 의한 데이터 수율의 악화로 나타난다. 이러한 문제는 HAPS(High Altitude Platform Station)를 사용하게 되면 해결이 된다. HAPS는 비행선을 이용한 시스템으로 LOS의 보장뿐만 아니라 기존의 지상망 이동통신에 접목해 좀 더 높은 성능을 낼 수 있다. 이 논문에서는 HAPS 시스템을 HSDPA 시스템과 결합한 시스템 모델을 제안하고 시뮬레이션을 통해 그 성능을 분석하고 가능성을 살펴본다. Today, there are many high speed data access systems that provide the truly “anytime and anywhere” services. Especially, HSDPA (High Speed Data Packet Access), one of the main third generation mobile communication systems, provides 14.4Mbps maximum data throughput. However, HSDPA will fail to provide high data throughput in hostile multipath fading environments due to lack of LOS (Line of Sight). HAPS (High Altitude Platform Station) is one of the solutions to this problem. HAPS system not only provides LOS, but it can also provide high data rate services to the conventional terrestrial systems. This paper proposes HAPS-HSDPA system model and compares performance of HSDPA and HAPS-HSDPA.

네트워크 침입탐지 기술 연구 동향

김동훈(Dong-Hoon Kim),손인수(Insoo Sohn) 대한전자공학회 2019 전자공학회논문지 Vol.56 No.8

최근 정보통신 기술이 발달하며 급격하게 정보량이 늘어나고, 그에 따라 해킹을 시도하는 사례가 많이 늘고 있다. 이에 따라 여러 종류의 공격 행위를 탐지하는 침입탐지 시스템이 필요하게 되었다. 침입탐지 시스템는 정해진 규칙(패턴)에 따라 시스템을 구성하는 오용 탐지(Misuse Detection)과 기존 데이터의 행동 유형을 ‘학습’하여 새로운 패턴의 데이터가 들어와도 판별할 수 있는 비정상 탐지(Anomaly Detection)가 있고, 최근에는 비정상 탐지를 기반에 둔 기계학습 기반 침입탐지 시스템을 연구한 논문이 많이 나오고 있다. 본 논문에서는 침입탐지 기술과 기계학습을 이용한 침입탐지 시스템에 관련된 논문들을 분석하여 침입탐지 시스템 생성 과정과 기계학습 기법의 종류에 따른 침입탐지 시스템 성능 평가를 수행한 연구 결과를 알아본다. 또한 SVM, NN, DT 기반 침입탐지 모델을 구현 및 생성하여 정확도, 탐지율, 오경보율을 측정하여 탐지 모델과 성능 측정 방법의 상관관계에 대해 알아보고 향후 기계학습 기반 침입탐지 시스템의 연구 방향을 제시한다. Recently, The amount of information increases as ICT develop and then Cracking more increases. For this reason, Intrusion Detection System that many kinds of Attacks are detected is needed. Intrusion Detection System is composed of two things. One is Misuse Detection that comprises system by fixed rule(Pattern) and the other is Anomaly Detection that differentiate the new Pattern of data by learning behavior type. There are many papers that studies Machine Learning based-on IDS that uses Anomaly Detection. In this paper, we analyze related papers that include Intrusion Detection Technology and Machine Learning based-on IDS. Using SVM, NN, DT, we create IDS and analyze the relationship between Detection Model and Performance measurement. and we suggest the way of research about Machine-Learning based-on IDS.

펨토셀 시스템 용량 개선을 위한 전력 제어 알고리즘

정재환(Jae-Hwan Jung),손인수(Insoo Sohn) 한국통신학회 2016 韓國通信學會論文誌 Vol.41 No.9