Hybrid Optical/RF Wireless Communication with Augmented Deep Learning for Indoor and Underwater IoT
- Abstract
- In this study, we propose advanced hybrid communication systems. Our goal is to ad- dress the limitations of the current Internet of Things technologies. These include chal- lenges in energy efficiency, long-range communication, and reliable connectivity. Our approach is tailored for challenging environments like underwater and indoor spaces. First, we investigate the feasibility of using radio frequency-based technology such as long-range (LoRA) in underwater-to-above-water wireless communication systems, as- sessing the link performance through theoretical and experimental approaches. The com- munication link quality is analyzed using key metrics such as received signal strength in- dicator, signal-to-noise ratio, and packet delivery ratio at different depths and distances, with various LoRa physical layer configurations. Based on the results, we designed a LoRa-based relay node to mitigate the limitations of underwater communication. Af- terward, we compared the system’s performance with and without the relay in terms of range, link quality, latency, and energy consumption. The results demonstrate that the relay significantly enhances communication performance without introducing additional power consumption or delay. Second, to further improve underwater communication, we design a hybrid Lo- Ra/optical relay node (LoOp), combining optical wireless communication and LoRa technologies to improve energy efficiency, communication range, and reliability. De- pending on the situation, either the LoRa side or the optical side of LoOp can serve as a transmitter or receiver. The experimental investigations reveal critical challenges for underwater LoRa applications, and we propose a link calibration strategy to opti- mize connectivity. In addition, we develop a data rate flow control mechanism to ensure smooth integration of LoRa and optical systems. We compared LoOp with another type of relay node to evaluate its performance in terms of delay and energy consumption. The experimental results show LoOp’s capabilities for underwater applications. Experimental observations and theoretical analysis demonstrate the benefits of the hybrid systems, prompting us to extend this approach to terrestrial applications by de- signing a visible light communication (VLC)-RF hybrid system tailored for indoor IoT networks. Therefore, we design and experimentally test a hybrid VLC/RF transceiver, featuring a channel sensing hardware/software subsystem. Utilizing this subsystem, we propose a link selection and switching system based on channel conditions. The hybrid nodes are integrated with sensors for indoor monitoring, and a bidirectional VLC/RF setup enables two-way communication. A multi-hop strategy is also applied to extend network coverage for wide-area indoor monitoring. To improve network reliability, we designed a hybrid access and network management protocol, compared against other protocols. To further enhance system performance, we propose an embedded intelli- gence scheme in the monitoring node for link switching and network configuration pur- poses. We compare the proposed hybrid bidirectional multi-hop communication system with other IoT connectivity technologies to evaluate its performance. The experimental results showcase the hybrid solution’s capabilities for enhancing IoT connectivity. Afterward, we present a VLC/LoRa network specifically designed for real-time, secure, and energy-efficient scenarios, such as fire detection, where immediate and secure data transmission is critical. The system employs a hybrid communication approach, integrating LoRa with VLC to mitigate the limitations of each technology when used independently. We designed and implemented a multi-hop network architecture where sensor data from monitoring nodes in individual rooms is transmitted to hybrid nodes located in a central corridor. These hybrid nodes then relay the data to a central gateway via VLC links. We experimentally evaluate the quality of both VLC and LoRa communication links. Based on the experimental results, the optimal configurations were selected and used to evaluate the proposed network performance. We embedded distributed machine learning directly into the monitoring nodes, enabling local inference and reducing latency and energy consumption. This also enhances data security by minimizing cloud transmissions and the risk of data interception. For this purpose, we experimentally tested various machine learning models on embedded devices, selecting the best one based on essential metrics such as accuracy, inference time, and power consumption. The results show that the designed hybrid network can provide a scalable, practical solution for IoT-based fire detection in complex indoor environments. Keywords Hybrid communication, IoT, LoRaWAN, Low-power networks, Machine learning, Optical wireless communication, Visible light communication, UIoT|이 연구는 현재 사물인터넷(IoT) 기술이 가진 에너지 효율성, 장거리 통신, 신뢰할 수 있는 연결성의 한계를
해결하기 위해 고급 하이브리드 통신 시스템을 제안하고 개발했다. 특히 실내와 수중 환경에서의 문제 해결을
목표로 했다.
먼저, 장거리(LoRa)와 같은 무선 주파수 기반 기술을 수중에서 수면 위로 연결되는 무선 통신 시스템에
사용할 수 있는지 조사하고, 이 링크의 성능을 이론적 및 실험적 접근을 통해 평가했다. 링크 품질은 RSSI, SNR,
PDR과 같은 주요 지표를 사용하여 다양한 깊이와 거리에서 LoRa 물리 계층 구성에 따라 분석했다. 이러한
결과를 바탕으로 수중 통신의 한계를 완화하기 위해 LoRa 기반 릴레이 노드를 설계했다. 이후, 릴레이 유무에
따른 시스템의 범위, 링크 품질, 지연 시간, 에너지 소비 측면에서 성능을 비교했다. 결과는 릴레이가 추가적인
전력 소비나 지연을 유발하지 않으면서 통신 성능을 크게 향상시킨다는 것을 보여주었다.
둘째로, 수중 통신을 더욱 개선하기 위해 에너지 효율성, 통신 범위, 신뢰성을 향상시키기 위한 하이브리드
LoRa/광학 릴레이 노드(LoOp)를 설계했다. 상황에 따라 LoOp의 LoRa 측 또는 광학 측이 송신기나 수신기로
사용될 수 있다. 실험적 연구는 수중 LoRa 응용 프로그램에서 중요한 과제를 제시하며, 연결성 최적화를 위한
링크 보정 전략을 제안하였다. 또한, LoRa와 광학 시스템의 원활한 통합을 위해 데이터 속도 흐름 제어 메커
니즘을 개발했다. LoOp와 다른 릴레이 노드를 비교하여 지연 및 에너지 소비 측면에서 성능을 평가했다. 실험
결과는 LoOp가 수중 응용에 적합함을 보여준다.
실험적 관찰과 이론적 분석을 통해 하이브리드 시스템의 장점을 입증했으며, 이를 실내 IoT 네트워크에 맞
춘 가시광 통신(VLC)-RF 하이브리드 시스템을 설계하여 지상 응용으로 확장하고자 했다. 하이브리드 VLC/RF
트랜시버를 설계하고 실험적으로 테스트했으며, 채널 감지 하드웨어/소프트웨어 하위 시스템을 통해 링크 선택
및 전환 시스템을 제안했다. 하이브리드 노드는 실내 모니터링을 위한 센서와 통합되어 양방향 VLC/RF 설정
이 가능하였다. 또한, 광범위한 실내 모니터링을 위해 다중 홉 전략을 적용했다. 네트워크의 신뢰성을 높이기
위해 하이브리드 액세스 및 네트워크 관리 프로토콜을 설계하고, 링크 전환 및 네트워크 구성을 위한 임베디드
인텔리전스 체계를 제안했다. 제안된 하이브리드 양방향 다중 홉 통신 시스템과 다른 IoT 연결 기술을 비교하여
성능을 평가한 결과, IoT 연결성 향상에 있어 하이브리드 솔루션의 역량이 확인되었다.
이후, 화재 감지와 같은 실시간, 보안, 에너지 효율 시나리오를 위해 특별히 설계된 VLC/LoRa 네트워크를
제시하였다. 이 시스템은 LoRa와 VLC를 통합하여 두 기술의 한계를 극복하는 하이브리드 통신 접근 방식을
채택했다. 개별 방의 모니터링 노드에서 수집된 센서 데이터를 중앙 복도에 위치한 하이브리드 노드로 전송하고,
이 하이브리드 노드가 데이터를 VLC 링크를 통해 중앙 게이트웨이로 중계하는 다중 홉 네트워크 아키텍처를
설계 및 구현했다. VLC와 LoRa 통신 링크의 품질을 실험적으로 평가하고, 최적의 구성을 선택하여 네트워크
성능을 평가했다. 모니터링 노드에 분산된 기계 학습을 직접 내장하여 지연 시간과 에너지 소비를 줄이고 데이터
보안을 향상시켰다. 이를 위해 다양한 기계 학습 모델을 임베디드 장치에서 실험적으로 테스트하고, 정확도,
추론 시간, 전력 소비를 기준으로 최적의 모델을 선택했다. 결과는 설계된 하이브리드 네트워크가 복잡한 실내
환경에서 IoT 기반 화재 감지에 대한 확장 가능하고 실용적인 솔루션을 제공할 수 있음을 보여준다.
- Author(s)
- BOLBOLI JAVAD
- Issued Date
- 2025
- Awarded Date
- 2025-02
- Type
- Dissertation
- Keyword
- Internet of Things, Wireless Communication, Machine Learning
- Publisher
- 국립부경대학교 대학원
- URI
- https://repository.pknu.ac.kr:8443/handle/2021.oak/33986
http://pknu.dcollection.net/common/orgView/200000860027
- Affiliation
- 국립부경대학교 대학원
- Department
- 대학원 인공지능융합학과
- Advisor
- Wan-Young Chung
- Table Of Contents
- 1 Introduction 1
1.1 How Does a Hybrid System Operate 4
1.2 Indoor Hybrid Wireless Communication Systems 5
1.2.1 Hybrid Systems with RF-Based WiFi, LoRa, and Small Cells 7
1.2.2 Hybrid Systems with Macrocells 7
1.2.3 Network Selection in Indoor RF/Optical Wireless Hybrid Systems 8
1.2.3.1 Uplink/Downlink Transmission 8
1.2.3.2 Traffic Type 8
1.2.3.3 Security Requirements 8
1.2.3.4 LOS vs. Non-Line-of-Sight (NLOS) 9
1.2.3.5 Illumination Requirements 9
1.2.4 Opportunities in Indoor Hybrid RF/Optical Wireless Systems 9
1.2.4.1 Traffic Offloading to Optical Networks 9
1.2.4.2 Enhanced Link Reliability 9
1.2.4.3 Seamless Mobility 10
1.2.4.4 Energy Efficiency 10
1.2.4.5 Security Enhancements 10
1.2.4.6 Interference Reduction 10
1.2.4.7 Spectral Efficiency 10
1.3 Underwater Hybrid wireless Communication Systems 11
1.3.1 Underwater Wireless Communication (UWC) Technologies and Applications 11
1.3.2 Challenges in Underwater Communication Technologies 11
1.3.2.1 Acoustic Communication 11
1.3.2.2 Underwater RF Communication 11
1.3.2.3 Underwater Optical Communication (UWOC) 12
1.3.3 Advantages and Limitations of UWOC 12
1.3.4 Hybrid Underwater Communication Systems and Advantages 12
1.3.4.1 RF/Optical Hybrid System 13
1.3.4.2 Acoustic/Optical Hybrid System 13
1.3.4.3 Enhanced Link Reliability 13
1.3.5 Criteria for Network Selection in UWC and Technical Consideration 13
1.3.5.1 Communication Distance 14
1.3.5.2 Traffic Type and Link Reliability 14
1.3.5.3 Technical Considerations 14
1.4 Hybrid (Mixed) RF/RF Communication 14
1.5 Problem Statement and Past Work 15
1.6 Thesis Contribution and Organization 16
2 LoRaWAN-assisted Relay for Underwater IoT 18
2.1 Introduction 18
2.2 Related Works 20
2.3 Electromagnetic Wave Propagation in Water 22
2.4 LoRa Channel Modeling 25
2.5 Experimental Setup and LoRa Frequency Selection 27
2.5.1 LoRa Configuration and Payload Length 29
2.5.2 Experimental Protocol 31
2.6 Results and Discussion 32
2.6.1 Model Validation of the Channel 32
2.6.2 Link Performance Analysis 33
2.6.3 Impact of LoRa Physical Parameters 35
2.7 Essence of Relay for Underwater communication 37
2.7.1 Challenge 1: Path Loss in Water and Air-water Boundary 37
2.7.2 Challenge 2: Blind Spot Phenomenon 39
2.7.3 Proposed Relay and Comparison With Direct Communication Link (UW-AW) 42
2.7.4 System Performance Comparison in the Presence of the Relay and Without the Relay 43
2.8 Impact of Environmental Parameters on the LoRa-based Communication System 48
2.9 Conclusion 51
3 Hybrid LoRa/Optical System for Underwater IoT 53
3.1 Introduction 53
3.2 Related Work and Motivation 55
3.3 Theory 56
3.3.1 EM Wave Propagation in Water 56
3.3.2 Link Budget Model 59
3.4 Experimental Design Process 60
3.4.1 LoRa Frequency and Configuration Selection 62
3.4.2 Communication at the Air-Water Interface 63
3.4.3 Underwater Communication 64
3.5 LoOp 67
3.5.1 Design Concept 67
3.5.2 Hardware Components 68
3.5.2.1 Optical Transmitter 68
3.5.2.2 Optical Receiver 68
3.5.2.3 LoRa Transmitter and Receiver 69
3.5.2.4 Microcontroller Unit 69
3.5.2.5 Power Management System 70
3.5.3 Physical Layer Communication Protocol 70
3.5.3.1 PDR Prediction for Link Performance Calibration 70
3.5.3.2 Data Rate Flow Control 71
3.6 Performance Evaluation 73
3.6.1 Communication Quality 73
3.6.2 Energy Consumption 77
3.6.3 Latency 81
3.7 Conclusion 82
4 Hybrid Protocol for Indoor Multi-Hop VLC/RF IoT Networks 84
4.1 Introduction 84
4.2 Intelligent Hybrid VLC/RF System Design 88
4.2.1 Main Subsystem 88
4.2.2 VLC Subsystem 88
4.2.3 RF Subsystem 89
4.2.4 Intelligent Switching Subsystem 90
4.2.4.1 Lower System 90
4.2.4.2 Upper System 91
4.2.5 Bidirectional Communication 91
4.2.6 Hybrid Multi-hop Communication 96
4.2.7 Hybrid Access and Network Management Protocol 97
4.3 Intelligent Scheme 100
4.3.1 Embedded Intelligent Scheme for Link Selection and Switching System 100
4.3.2 Model Deployment on the Embedded Device 101
4.4 Result and Discussion 102
4.4.1 The physical layer communication quality and configuration selection experiments 102
4.4.2 Bidirectional and Multi-hop Communication Experiment 107
4.4.3 Latency 108
4.4.4 Energy Consumption 110
4.5 Conclusion 112
5 Distributed Embedded ML-Driven Hybrid VLC/LoRa Network for Indoor IoT 113
5.1 Introduction 113
5.2 Related Work 115
5.3 System Design 117
5.3.1 Hybrid VLC/LoRa Nodes 117
5.3.1.1 Main Systen 117
5.3.1.2 VLC Transmitter and Receiver 117
5.3.1.3 LoRa Modulation and Transmitter and Receiver Selection 119
5.3.1.4 Power Management Unit 119
5.3.2 Monitoring Nodes Design 119
5.4 Network Design 120
5.4.1 Network Architecture 120
5.4.2 Multi-hop Communication 120
5.4.3 Distributed Embedded Intelligence Approach 121
5.5 Performance Evaluation 122
5.5.1 Communication Quality 122
5.5.2 Model Selection and Evaluation 126
5.5.3 System Energy Consumption 130
5.5.4 Network Latency 132
5.6 Conclusion 133
6 Conclusion 136
7 Future Works 137
Bibliography 139
논문요약 157
Acknowledgement 159
Publications Based on the Thesis 160
- Degree
- Doctor
-
Appears in Collections:
- 대학원 > 인공지능융합학과
- Authorize & License
-
- Authorize공개
- Embargo2025-12-01
- Files in This Item:
-
Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.