Myongji University · DAN Lab · Integrated MS–PhD 명지대학교 · DAN 연구실 · 석·박사 통합과정

Jiha Kim 김지하

I study reliable wireless systems and safety intelligence for Physical AI under real-world constraints.

실제 환경의 제약 안에서 신뢰할 수 있는 무선 시스템과 Physical AI 안전지능을 연구합니다.

My work connects IEEE 802.11 multi-link MAC, constrained reinforcement learning, conformal OOD detection on edge NPUs, and robust visual perception. I focus on systems whose risk can be measured, explained, and verified beyond benchmark averages.

IEEE 802.11 멀티링크 MAC, 제약 기반 강화학습, 엣지 NPU 기반 컨포멀 OOD 탐지, 강건한 시각 인지를 연결합니다. 평균 성능을 넘어 위험을 측정하고 설명하며 검증할 수 있는 시스템을 만드는 데 관심이 있습니다.

Position 과정 Integrated MS–PhD Candidate, 2020–present 석·박사 통합과정, 2020–현재

Dept. 학과 Information & Communication Engineering 정보통신공학과

Lab 연구실 DAN Lab — Distributed AI and Networked Robotics — 분산 AI 및 네트워크 로보틱스

Advisor 지도교수 Prof. Hyunhee Park 박현희 교수

Email yaki5896@mju.ac.kr

ORCID 0000-0002-6711-6812

Elsewhere 링크 Google Scholar · GitHub · Blog

10 Published papers게재 논문

17 Domestic국내 발표

4 In review / revision심사·수정 중

7 Patents filed특허 출원

63 Citations피인용 Google Scholar

5 h-index i10 · 1

Google Scholar · 누적 지표 · 마지막 확인 2026-11-13All-time metrics · last checked 2026-11-13

Research연구 분야

Reliability-oriented MAC for Wi-Fi 8 Wi-Fi 8 신뢰성 중심 MAC

Multi-link contention resists closed-form analysis once the EMLSR single-active-radio constraint couples queues across bands. I build Markov/fixed-point closures that stay analytically auditable, then drive contention windows from MAC observables inside the standard EDCA loop.

EMLSR의 단일 활성 라디오 제약이 대역 간 큐를 결합시키면 멀티링크 경쟁은 닫힌 형태로 풀리지 않습니다. 분석 가능성을 유지하는 Markov/고정점 폐형을 세우고, 표준 EDCA 루프 안에서 MAC 관측값으로 경쟁 윈도우를 제어합니다.

802.11bnMLO / EMLSRMarkov

Constrained RL for network control 제약 기반 강화학습 네트워크 제어

Ultra-high reliability is a tail problem, not a mean problem. I use Lyapunov drift-plus-penalty to put p99 and p99.9 violation rates directly into the learning signal, with federated aggregation weighted by measured interference rather than plain averaging.

초고신뢰성은 평균이 아니라 꼬리 분위의 문제입니다. Lyapunov drift-plus-penalty로 p99·p99.9 위반율을 학습 신호에 직접 넣고, 단순 평균 대신 측정된 간섭에 비례한 가중치로 연합 집성을 수행합니다.

MARLLyapunovFederated

Conformal safety intelligence for Physical AI Physical AI를 위한 컨포멀 안전지능

My doctoral work builds a Physical Safety Token with conformal risk control, edge-NPU inference, and statistical Sim-to-Real verification. It begins with robot manipulation and extends to cascading physical and communication anomalies in UAV base stations.

학위 연구는 컨포멀 위험 제어, 엣지 NPU 추론, 통계적 Sim-to-Real 검증을 결합한 물리 안전 토큰을 설계합니다. 로봇 조작을 시작으로 UAV 기지국에서 연쇄되는 물리·통신 이상 탐지까지 확장합니다.

PST-CPEdge NPUSim-to-Real

Robust video analysis and understanding 강건한 영상 분석 및 이해

I study action recognition and early anomaly detection under real deployment constraints. The work separates appearance from motion, fuses multi-scale spatio-temporal evidence, and reuses motion cues for frame-level localization and anticipation while accounting for edge-compute efficiency.

실제 배포 환경의 제약 아래 행동 인식과 조기 이상 탐지를 연구합니다. 외형과 동작을 분리하고 다중 스케일 시공간 증거를 융합하며, 동작 단서를 프레임 단위 위치 추정과 사건 예측에 재사용합니다. 엣지 실행을 고려해 정확도와 연산 효율을 함께 평가합니다.

Action RecognitionVideo AnomalyEdge Vision

Trajectory — where the work has gone 연구 궤적 — 주제의 이동

Blockchain & distributed ledgers 블록체인 · 분산 원장 IMIS

11편

Explainable AI & GAN 설명 가능한 AI · GAN ICUFN · IMIS · ICT Express

4 papers4편

Reinforcement learning for wireless resource control 무선 자원 제어를 위한 강화학습 ICTC · IEEE Access

2 papers2편

Wi-Fi multi-link & ultra-high reliability Wi-Fi 멀티링크 · 초고신뢰성 BWCCA · KICS · IEEE TMC · TCCN · TNSE

7 papers7편

Video understanding 비디오 이해 IMIS · JCCI · IEEE JSAS · IEEE TIP

4 papers4편

Physical AI safety intelligence Physical AI 안전지능 PST-CP · Edge NPU · Sim-to-Real

PM책임

Publications논문

10+04

Output by year연도별 실적 국제 10편 · 국내 17편 · 심사·수정 중 4편10 international · 17 domestic · 4 in review / revision

국제 저널 · 3편Global journal · 3
국제학회 · 7편International conf. · 7
국내학회 · 17편Domestic conf. · 17
심사·수정 중 · 4편Review / revision · 4

마크 1개 = 논문 1편. 사이트 목록에서 자동 집계합니다. 국제 실적은 게재 확정을 포함하며, 국내 발표는 제1저자·공저 논문을 모두 포함합니다.

One mark = one paper, counted from this site's records. International papers include accepted work; domestic presentations include first- and co-authored work.

Global Journal국제 저널

2026Journal

ASTRA: Adaptive Spatio-Temporal Representation for Action Recognition via Importance-Guided Motion Learning

J. Kim, S. Park, and H. Park

IEEE Journal of Selected Areas in Sensors — accepted, to appear — 게재 확정

Abstract초록

3D-CNN action recognizers learn appearance and motion in an entangled way, so they lean on static visual cues instead of true motion — which hurts generalization and interpretability on resource-constrained AIoT edge devices. ASTRA is a dual-pathway architecture that separates semantic appearance from pure motion through three components: an STFormer module that models space and time adaptively through cross-attention and estimates frame-wise importance; Importance-Guided Temporal Perturbation, which isolates appearance-free motion by comparing original and perturbed inputs; and a Motion Gated Unit that refines the motion features through dual-gate modulation. Under a leak-free, video-level protocol ASTRA reaches 97.88% on UCF-101 and 81.05% on HMDB-51 with an MViT-v2-S backbone. Against its own R(2+1)D-18 backbone, the motion pathway adds a significant +2.81%p on HMDB-51 (McNemar p < 0.001) while staying equivalent on appearance-saturated UCF-101, at 32.9M parameters, 41.9 GFLOPs and 8.2 ms GPU latency.

UCF-101 97.88% HMDB-51 81.05% HMDB-51 +2.81%p · p < 0.001 32.9M params · 8.2 ms

2024Journal

Dynamic Transmission and Delay Optimization Random Access for Reduced Power Consumption

J. Kim, Y. Kim, S. Park, and H. Park

IEEE Access, vol. 12, pp. 55033–55050, April 2024

Abstract초록

Massive machine-type communications congest the random-access channel and burn power on repeated preamble transmissions. DTDO-RA adjusts the backoff indicator (BI) according to the number of transmissions, and uses reinforcement learning — Q-learning and DDPG — to tune the BI together with the maximum number of preamble transmissions (Max Tx) as the network state changes. Widening the backoff range under heavy traffic keeps the procedure succeeding: the success rate, which falls at 10,000 UEs under the standard procedure, holds until 15,000 UEs. At 15,000 UEs a larger BI cuts transmit power by about 98% (3.71 to 0.072 mJ) at the cost of longer access delay, while letting Max Tx widen the backoff range lowers the delay by up to 4.3× relative to BI adjustment. Simulations run to 50,000 UEs and also show that shrinking Max Tx too far raises power rather than saving it.

tx power −98.06% at 15k UEs success-rate knee 10k → 15k UEs 4.3× lower delay than BI tuning Q-learning · DDPG · 50k UEs

2023Journal

Limited Discriminator GAN using explainable AI model for overfitting problem

J. Kim and H. Park

ICT Express, vol. 9, no. 2, pp. 241–246, 2023

Abstract초록

GANs are widely used to augment scarce data, but a discriminator that leans too hard on the training set overfits, and the generator ends up reproducing images that look like the training images — augmentation that adds nothing. LDGAN opens the discriminator with LIME to show which image regions drive its real/fake decision, and finds that an unconstrained discriminator relies on the whole image, background included. It then limits the discriminator's training with a threshold, set either on the discriminator's mean loss (about 0.5) or on the generator's loss, so that the decision rests on the object rather than the background and the generated images stay diverse. Of 500 images generated per model, 254 from DCGAN show usable eye and nose regions, against 324 and 360 under the two LDGAN limits.

eye/nose regions in 360/500 (DCGAN 254) D-loss limit 324/500 D-loss threshold 0.5 LIME · CIFAR-10 · Dogs vs. Cats

International Conferences국제 학술대회

2025IMIS

STCA-Net: Spatio-Temporal Convolutional Attention Network for Efficient Action Recognition in Videos

J. Kim and H. Park

International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS), 2025

2024BWCCA

Multi-Link/Multi-AP Coordination Based Joint Transmission for Seamless Roaming in IEEE 802.11bn (Wi-Fi 8)

J. Kim and H. Park

International Conference on Broadband and Wireless Computing, Communication and Applications (BWCCA), 2024

2023ICTC

Optimization of Reinforcement Learning-Based Backoff Indicator for 5G NR Random Access Procedure

J. Kim, C. You, and H. Park

International Conference on ICT Convergence (ICTC), 2023

2022IMIS

Hierarchical output model of CNN learning using multi-label datasets

J. Kim, A. A. José, J. Kim, Y. Kim, and H. Park

International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS), 2022

2021ICUFN

OA-GAN: Overfitting Avoidance Method of GAN Oversampling based on Explainable AI

J. Kim and H. Park

International Conference on Ubiquitous and Future Networks (ICUFN), July 2021

2021IMIS

Reduced CNN Model for Face Image Detection with GAN Oversampling

J. Kim and H. Park

International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS), 2021

2020IMIS

An Accounting Ledger System using the Hyperledger Fabric-based Blockchain

J. Kim, H. W. Kim, W. S. Hyun, and H. D. J. Jeong

International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS), 2020

Under Review & Revision심사·수정 중

2026 In review심사 중

Reliability-Aware Adaptive Contention Control for EMLSR Multi-Link Operation in IEEE 802.11bn WLANs

J. Kim, S. Park, and H. Park

IEEE Transactions on Mobile Computing — under review — 심사 중

Abstract초록

IEEE 802.11bn Ultra High Reliability (UHR) extends multi-link operation toward a reliability-oriented MAC, but multi-link contention under the EMLSR active-link constraint resists closed-form analysis. We develop a Markov/fixed-point closure for three-band MLDs coexisting with single-link legacy stations: exact under saturated, per-link-independent STR; reducing to an availability-based approximation under saturated EMLSR with an explicit validity range; and extending to unsaturated traffic through per-link Bernoulli arrivals. On top of it we build a controller that adapts the per-link minimum contention window from MAC observables inside the standard EDCA/EMLSR loop. Against SLD, STR-MLD and static-EMLSR baselines it raises throughput by 10.6%, lowers 95th-percentile HOL-service delay by 28%, and improves Jain's fairness from 0.910 to 0.954 at 48 stations. Triangulation across the analytical model, a discrete-event simulator and ns-3 closes within a 2–4% mean relative gap.

+10.6% throughput −28% p95 HOL delay fairness 0.910 → 0.954 model ↔ ns-3 gap 2–4%

2026 Revision수정 중

LyMAPPO: Lyapunov-Constrained Independent Learning for Ultra-High Reliability in Dense Multi-Band Wi-Fi

J. Kim, S. Park, and H. Park

IEEE Transactions on Cognitive Communications and Networking — revision — 수정 중

Abstract초록

IEEE 802.11bn defines ultra-high reliability through per-BSS deadline-violation-rate targets. We study these constraints in dense multi-band deployments with heterogeneous AP link sets. LyMAPPO combines independent per-AP PPO with drop-inclusive rate-form Lyapunov duals, one dual-derived congestion price per beacon for cross-BSS coordination, and an auxiliary lower-tail quantile critic. Training and evaluation use an ns-3-in-the-loop environment with 16 APs, three bands, shared-buffer multi-link operation and Markov channels. Across 16 seeds with 110 s episodes, LyMAPPO achieves 15.1/16 feasible BSSs and a network p99 of 0.0017 — 6.1× to 25× lower than three static baselines. The drift-plus-penalty analysis accounts for clipping and establishes a run-time dual-stability condition under the Slater assumption.

network p99 0.0017 6.1–25× lower than static 15.1/16 feasible BSS 16 APs · 3 bands · ns-3

2026 In review심사 중

STCA-ViT: Spatio-Temporal Cross-Attention Vision Transformer for Video Action Recognition

J. Kim, S. Park, and H. Park

IEEE Transactions on Image Processing — under review — 심사 중

Abstract초록

STCA-ViT is a hybrid framework integrating a VideoMAE ViT-Base backbone with bidirectional spatio-temporal cross-attention, Sobel-based boundary detection for motion gating, and YOLO-based person attention in parallel. It adds residual contribution fusion with gradient reversal for feature diversity, orthogonal regularization, a multi-head ensemble classifier, and an optional weakly-supervised temporal action localization head that recycles motion features into frame-level scores without temporal annotations. STCA-ViT reaches 90.46% Top-1 on HMDB-51 — above VideoMAE V2 ViT-g (86.08%) — with 98.29% on UCF-101, 79.44% on Kinetics-400 and 65.01% on Something-Something-v2, using 110M parameters and 305 GFLOPs.

HMDB-51 90.46% UCF-101 98.29% K-400 79.44% 110M params · 305 GFLOPs Code ↗코드 ↗

2026 In review심사 중

Toward Wi-Fi 9 for Physical AI: A Task-Centric Survey of Local Intelligence Infrastructure

IEEE Transactions on Network Science and Engineering — under review — 심사 중

Domestic국내 발표

Korean Institute of Communications and Information Sciences (KICS) conferences and JCCI. Titles are given as published, in Korean.

한국통신학회(KICS) 종합학술발표회 및 통신정보 합동학술대회(JCCI) 발표 논문입니다.

Domestic — first author국내 발표 — 제1저자

2026KICS Summer하계통신

IW-FedMAPPO: 비대칭 OBSS Wi-Fi 8 네트워크에서 초고신뢰 조정 MLO 스케줄링을 위한 간섭 가중 연합 다중 에이전트 근접 정책 최적화

Interference-Weighted Federated Multi-Agent PPO for Ultra-Reliable Coordinated MLO Scheduling

김지하, 박현희*

2026JCCIJCCI

SEAR: 공유적 증거 이상 및 위험 탐지 — 비디오 탐지 및 다중 시간 지평 예측 통합 프레임워크

Shared Evidential Anomaly and Risk Detection

김지하, 박현희*

2026KICS Winter동계통신

QA-FedDQN: 독립 배치 다중 AP Wi-Fi 네트워크를 위한 품질 인식 적응적 연합 심층 Q-Network

Quality-Aware Adaptive Federated Deep Q-Network for Distributed Multi-AP Wi-Fi

김지하, 박현희*

2025KICS Summer하계통신

PRIME Wi-Fi: Wi-Fi의 멀티링크 효율성을 위한 예측적 리소스 인텔리전스

Predictive Resource Intelligence for Multi-Link Efficiency in Wi-Fi

김지하, 박현희*

2024KICS Fall추계통신

동적 컨볼루션과 다중 스케일 시공간 주의 메커니즘을 사용한 비디오 분류 모델

Video classification with dynamic convolution and multi-scale spatio-temporal attention

김지하, 박현희*

2024JCCIJCCI

자율주행을 위한 Unknown Object Detection 연구 동향: Multi-Modal 접근과 불확실성 인식 기법

Research trends in unknown object detection for autonomous driving

김지하, 김용호, 김건희, 김 환, 박현희*

2023KICS Summer하계통신

수직 및 수평 필터를 활용한 이중 생성자 구조를 통한 GAN의 이미지 생성 품질 향상

Improving GAN image quality with a dual-generator structure

김지하, 박현희*

2023KICS Winter동계통신

5G mMTC 환경에서 2-pair UE NOMA를 이용한 sectorization 랜덤 액세스 절차의 성능 비교

Sectorized random access with 2-pair UE NOMA in 5G mMTC

김지하, 박소민, 김용호, 유철우, 박현희*

2022JCCIJCCI

다중 클래스 데이터 셋에서의 딥 러닝 성능 개선을 위한 다중 출력 모델 제안

A multi-output model for deep learning on multi-class datasets

김지하, 김지나, 김용호, 박현희*

2022KICS Winter동계통신

Multi-Link Operation의 Internal Interference를 위한 적응적 Contention Window 기법

Adaptive contention window for internal interference in multi-link operation

김지하, 김병찬, 박현희*

2021KICS Summer하계통신

GAN Data augmentation을 위한 검증 및 설명 가능한 VX-GAN 모델

VX-GAN: a validation and explainable GAN model for data augmentation

김지하, 박현희*

2021JCCIJCCI

CNN 알고리즘을 이용한 소형 임베디드 인물 판별 시스템 모델

Face recognition model for small embedded systems based on CNN

김지하, 박현희*

Domestic — co-author국내 발표 — 공저

2023KICS Fall추계통신

UAV를 활용한 통합 시맨틱 객체 인식 기반 도로 밀집도 이상 영역 탐지

Road-density anomaly detection from UAV semantic object recognition

김 환, 김지하, 박현희*

2021KICS Fall추계통신

SDN 환경에서 서비스별 네트워크 요구사항에 따른 우선순위 결정 기법

Service-aware prioritization by network requirement in SDN

우기문, 김형빈, 김지하, 박현희*

2021KICS Summer하계통신

사용자별 가중치 표준 편차를 활용한 Federated Learning 성능 향상 기법

FedSD: federated learning using per-user weight standard deviation

김형빈, 김용호, 우기문, 김지하, 박현희*

2021KICS Summer하계통신

NHANES 데이터를 이용한 생활 패턴 기반의 우울증 예측 머신러닝 기법

Lifestyle-based depression prediction from NHANES data

김용호, 김지하, 우기문, 김형빈, 박현희*

2021KICS Summer하계통신

MEC 환경에서 5G-V2X 기반 사고 예방 시스템 제안

A 5G-V2X accident-prevention system in MEC environments

우기문, 김형빈, 김용호, 김지하, 박현희*

Patents특허

Korean patent applications filed through the Myongji University Industry–Academic Cooperation Foundation, with Prof. Hyunhee Park.

명지대학교 산학협력단을 통해 박현희 교수와 공동 출원한 국내 특허입니다.

2024

Video Classification System and Method Using Dynamic Convolution and a Multi-Scale Spatio-Temporal Attention Mechanism

10-2024-0198827 · published KR 10-2026-0104813 A, 6 Jul 2026공개 KR 10-2026-0104813 A, 2026.07.06

2024

A Method for Performing Handover of Non-AP MLD for Multiple AP MLDs Performing Joint Transmission, and a Device Therefor

10-2024-0194211 · published KR 10-2026-0101234 A, 30 Jun 2026공개 KR 10-2026-0101234 A, 2026.06.30

2024

A Method for Performing Cooperative Transmission of UMAC-based AP MLD in Wireless LAN Network and an Apparatus Therefor

10-2024-0194212 · published KR 10-2026-0101235 A, 30 Jun 2026공개 KR 10-2026-0101235 A, 2026.06.30

2024

A Method for Receiving Cooperative Transmission Data of Non-AP MLD in Wireless LAN Network and an Apparatus Therefor

10-2024-0194213 · published KR 10-2026-0101236 A, 30 Jun 2026공개 KR 10-2026-0101236 A, 2026.06.30

2024

A Method for Parameter Exchange for Joint Transmission Among Multiple AP-MLDs Supporting Multi-Link, and Device Therefor

10-2024-0194214 · published KR 10-2026-0101237 A, 30 Jun 2026공개 KR 10-2026-0101237 A, 2026.06.30

2023

Method for Performing Random Access Procedure Based on Reinforcement Learning

10-2023-0196634 · published KR 10-2025-0104336 A, 8 Jul 2025공개 KR 10-2025-0104336 A, 2025.07.08

2022

Method and Apparatus for GAN-based Model Learning

10-2022-0024885 · published KR 10-2023-0127508 A, 1 Sep 2023공개 KR 10-2023-0127508 A, 2023.09.01

Projects연구 과제

국가연구개발 과제 8개, 산학 연구 6개, 표준특허 지원사업 1개와 응용 연구 1개. 무선 네트워크부터 영상 이해, 임베디드 AI, Physical AI 안전지능까지 연구와 산업 현장을 연결합니다.

Eight national R&D projects, six industry research projects, one standard-essential patent support program, and one applied research topic—connecting wireless systems, video understanding, embedded AI, and Physical AI safety.

Recent national R&D participation · 2021–2027 최근 국가 R&D 참여 이력 · 2021–2027

PST-CP · PM2026.09–2027.02

시맨틱 통신·제어Semantic communication & control2026.08–2027.02

Flying BS · AI-MAC2024.03–2027.02

Explainable AI · WLAN2021.04–2025.12

6G STAR-MAC2021.04–2023.12

실감 콘텐츠·홀로그램Immersive content & holograms2022.04–2023.12

VR·AR 콘텐츠 생성VR/AR content generation2022.01–2022.12

막대는 참여목록에 기록된 참여기간입니다(향후 예정기간 포함). 연차별 기록은 통합했으며, 전체 과제기간과는 다릅니다. 진한 막대는 연구책임자(PM), 파란 막대는 참여연구원입니다.

Bars show recorded participation, including scheduled future periods, with annual records merged—not the full project duration. Dark: principal investigator / PM. Blue: participating researcher.

국가연구개발 과제National R&D projects8

2026.09–
2027.02 NRF · PM

Conformal Physical Safety Token on NPU for Safety Intelligence in Physical AI and Communication Systems: Coverage-Guaranteed OOD Detection and Sim-to-Real Verification Physical AI 및 통신 시스템 안전지능을 위한 NPU 기반 컨포멀 물리 안전 토큰: 커버리지 보장 OOD 검출과 Sim-to-Real 자동 검증

2026 Doctoral Student Research Grant · Principal investigator / PM: Jiha Kim 2026년도 박사과정생연구장려금지원사업 · 연구책임자(PM): 김지하

Develops a conformal Physical Safety Token (PST-CP) that controls false-alarm risk under stated data conditions, runs in real time on an edge NPU, and diagnoses Sim-to-Real gaps. The core robot-arm study extends to joint physical and communication OOD detection for UAV base stations.

명시한 데이터 조건에서 오경보 위험을 제어하는 컨포멀 물리 안전 토큰(PST-CP)을 설계하고, 엣지 NPU 실시간 추론과 Sim-to-Real 갭 자동 진단을 통합합니다. 로봇 팔 중심의 핵심 연구를 UAV 기지국의 물리·통신 이중 OOD 탐지로 확장합니다.

Physical AIPST-CPEdge NPUSim-to-RealUAV-BS

수행기관Institution: 명지대학교 · 한국연구재단(NRF)Myongji University · NRF
전체 과제Full project: 2026.09–2028.08
등록 참여Recorded participation: 2026.09–2027.02 (1차년도)(year 1)

2026.08–2027.02NRF

동영상 기반 객체 제어를 위한 시맨틱 통신-제어 시스템 통합 설계 Semantic Communication-Control Co-Design for Video-based Object Control

성균관대학교 · 한국연구재단Sungkyunkwan University · NRF

동영상에서 추출한 객체 상태와 시맨틱 통신을 연결하고, 원격 객체 제어를 위한 통신·제어 시스템을 함께 설계하는 과제입니다.

A project integrating video-derived object states with semantic communication and joint communication–control design for remote object control.

역할Role: 참여연구원Participating researcher
참여기간Participation: 2026.08–2027.02
전체 과제Full project: 2025.03–2030.02

2024.03–2027.02NRF

AI-MAC Protocol for Intelligent Flying Base Stations Based on Distributed Machine Learning 분산 머신러닝 기반 지능형 플라잉 기지국을 위한 AI-MAC 프로토콜

명지대학교 · 한국연구재단Myongji University · NRF
2022R1A2C2005705

Distributed learning for MAC-layer decisions in UAV-assisted networks; the origin of my multi-UAV energy and Age-of-Information work.

UAV 기반 네트워크의 MAC 계층 의사결정을 위한 분산 학습 연구. 다중 UAV 에너지·AoI 최적화 연구의 출발점입니다.

역할Role: 참여연구원Participating researcher
참여기간Participation: 2024.03–2027.02
전체 과제Full project: 2022.03–2027.02

2022.04–2023.12KOCCA

미세 패터닝 기술을 활용한 실감 콘텐츠 복제 방지 및 홀로그램 활용 기술 개발Micro-patterning for Immersive Content Copy Protection and Hologram Applications

한국전자기술연구원 · 한국콘텐츠진흥원Korea Electronics Technology Institute · KOCCA

미세 패터닝을 활용한 실감 콘텐츠의 복제 방지 및 홀로그램 활용 기술 개발 과제입니다.

Development of micro-patterning technologies for immersive content copy protection and hologram applications.

역할Role: 참여연구원Participating researcher
참여기간Participation: 2022.04–2023.12
전체 과제Full project: 2022.04–2024.12

2022.01–2022.12IITP

이용 환경을 반영하는 자율적 VR·AR 콘텐츠 생성 기술개발Development of Autonomous VR and AR Content Generation Technology Reflecting Usage Environment

서울과학기술대학교 산학협력단 · 정보통신기획평가원Seoul National University of Science and Technology · IITP
2020-0-00994

이용 환경을 반영하여 VR·AR 콘텐츠를 자율적으로 생성하는 기술 개발 과제입니다.

Development of autonomous VR/AR content generation adapted to the usage environment.

역할Role: 참여연구원Participating researcher
참여기간Participation: 2022.01–2022.12
전체 과제Full project: 2020.04–2022.12

2021.04–2025.12IITP

설명가능한 인공지능 기반 무선랜 네트워크 시스템 고도화 핵심 기술 연구Research on Advanced Core Technologies for WLAN Based on Explainable AI

명지대학교 · 정보통신기획평가원Myongji University · IITP
2021-0-00990

설명가능한 인공지능을 무선랜 네트워크에 적용하여 네트워크 지능화와 다중 AP 운용을 위한 핵심 기술을 연구하는 과제입니다.

Research into explainable AI for intelligent WLAN systems and core technologies for multi-AP operation.

역할Role: 참여연구원Participating researcher
참여기간Participation: 2021.04–2025.12
전체 과제Full project: 2021.04–2025.12

2021.04–2023.12IITP

6G 서비스를 위한 인공지능/머신러닝 기반 자율형 MAC 개발Development of the 6G Service Targeted AI/ML-based Autonomous-Regulating Medium Access Control (6G STAR-MAC)

한양대학교 · 정보통신기획평가원Hanyang University · IITP
2021-0-00368

6G 서비스에 대응하는 AI/ML 기반 자율형 매체 접근 제어와 적응적 무선 자원 할당 기술을 개발하는 과제입니다.

Development of AI/ML-driven autonomous medium access control and adaptive radio resource allocation for 6G services.

역할Role: 참여연구원Participating researcher
참여기간Participation: 2021.04–2023.12
전체 과제Full project: 2021.04–2025.12

2017.12–2018.11TIPA

클라우드 기반의 인공지능 검색광고 플랫폼 개발Cloud-Based AI Search Advertising Platform

이글루커뮤니케이션즈 · 중소기업기술정보진흥원Igloo Communications · TIPA

첫걸음기술개발사업의 검색광고 플랫폼 개발 과제. NTIS 본인 이력에 등록된 2017·2018년 참여기록을 하나로 통합했습니다.

A search advertising platform project under the First-Step Technology Development program. Overlapping 2017 and 2018 personal NTIS participation records are consolidated.

역할Role: 참여연구원Participating researcher
참여기간Participation: 2017.12.01–2018.11.30
과제번호Project reference: C0563763

2026.10.07. NTIS 본인 참여이력 조회와 PDF·Excel 대조 기준입니다. 참여기간은 등록된 예정기간을 포함하며, 연차별 과제는 하나로 통합했습니다. 영문명이 없는 과제는 번역했습니다.

Based on personal NTIS participation records checked on 7 October 2026 and cross-checked against PDF/Excel records. Dates include registered future periods; annual entries are consolidated. Titles without an official English version are translated.

산학 연구 · 이노피아테크Industry research · Innopia6

개인 이름이 기재된 연구인력 명단·보고서와 본인 업무 기록을 바탕으로 정리했습니다. 표시 연도와 일자는 확인한 활동자료의 날짜이며, 공식 개인 참여기간은 별도 확인 중입니다. 과제 계약기간은 개인 참여기간과 구분합니다.

Based on named research-team records, reports, and authored project correspondence. Years and dates identify documented activity; formal individual participation periods remain to be confirmed. Project contract dates are listed separately.

2025활동자료 기준Documented activity