Dokumentasi Proyek

Perancangan Desain Jaringan Smart Campus IoT Network Terpusat dengan Implementasi AI Agent Berbasis LLM untuk Otomatisasi Troubleshooting Gateway via API

Proyek ini mengintegrasikan jaringan IoT kampus cerdas dengan kecerdasan buatan berbasis Large Language Model (LLM) untuk mengotomatiskan proses pemantauan, deteksi anomali, dan troubleshooting gateway secara real-time melalui antarmuka API terpusat.

LLM-powered AI Agent Real-time IoT Monitoring Automated Troubleshooting Smart Gateway Management REST / WebSocket API Edge & Cloud Computing
0
IoT Devices
0
Gateways
0
API Endpoints
0
% Uptime Target

Tim Peneliti

Mahasiswa dan dosen pembimbing yang terlibat dalam pengembangan proyek ini.

AR
Ahmad Rizky Maulana
NIM: 2021310001
Project Lead

Network Architecture & System Integration

FN
Farhan Nur Hidayat
NIM: 2021310047
Network Engineer

IoT Infrastructure & Gateway Configuration

DS
Dinda Sari Pertiwi
NIM: 2021310082
AI Developer

LLM Integration & Intelligent Automation

RW
Rizal Wahyu Pratama
NIM: 2021310115
Frontend Developer

Dashboard UI, API Design & Documentation

MH
Dr. Muhammad Hakim, S.T., M.T.
NIDN: 0012037802
Dosen Pembimbing

Pakar Jaringan Komputer & Kecerdasan Buatan
Departemen Teknik Telekomunikasi

Arsitektur Sistem

Diagram arsitektur berlapis sistem Smart Campus IoT Network dengan AI Engine terpusat.

☁ CLOUD / CENTRAL SERVER ⚡ EDGE / FOG COMPUTING 🔌 GATEWAY LAYER 📡 IoT DEVICES Central Server InfluxDB · Grafana Docker · Kubernetes Web Dashboard React · Bootstrap D3.js · WebSocket Message Broker MQTT Broker Mosquitto · RabbitMQ 🤖 AI Engine LLM Agent (GPT/LLaMA) TensorFlow · PyTorch Anomaly Detection API Gateway Node.js · Nginx · REST Auth · Rate Limit Edge Node #1 Raspberry Pi 4 Local Processing Edge Node #2 Jetson Nano ML Inference Fog Computing Data Aggregation Preprocessing · Filter Gateway A LoRaWAN · MQTT Gedung A–C Gateway B Zigbee · CoAP Gedung D–F Gateway C WiFi · BLE Outdoor Campus Gateway D 5G NR Parking / Labs 🌡 Env Sensor Temp · Humidity CO₂ · AQI ⚡ Smart Meter Energy · Power Voltage · Current 📷 CCTV IP Camera 4K RTSP Stream 🔒 Access Ctrl RFID · Biometric Door Lock 💡 Smart Light Dimmer · Lux Motion Trigger 🚗 Parking Ultrasonic LiDAR · Infrared 🌊 Water Meter Flow · Level 📶 WiFi AP 802.11ax 🔥 Fire Sensor Smoke · Heat 🌿 Greenhouse Soil · Plant + More Devices (expandable) Plug & Play · Auto Discovery Legend: Wireless / Sensor Data Gateway Uplink Edge → Cloud AI ↔ Cloud Interaction

Technology Stack

Teknologi dan framework yang digunakan dalam pengembangan sistem ini.

Node.js
v20 LTS
Backend API server & real-time WebSocket events
Python
v3.11
AI/ML pipeline, data processing & automation scripts
TensorFlow
v2.14
LLM model serving & anomaly detection neural nets
PyTorch
v2.1
Fine-tuning LLaMA & custom transformer models
MQTT
v5.0
Lightweight IoT messaging protocol, QoS 0/1/2
CoAP
RFC 7252
Constrained Application Protocol for low-power devices
InfluxDB
v2.7
Time-series database for IoT sensor telemetry storage
Grafana
v10.2
Observability dashboards & alerting for time-series data
Docker
v24
Containerization for consistent service deployment
Kubernetes
v1.28
Container orchestration, auto-scaling & self-healing
Nginx
v1.25
Reverse proxy, load balancer & API gateway frontend
D3.js
v7.8
Interactive SVG visualizations, topology & charts

Timeline Implementasi

Rencana pelaksanaan proyek selama 6 bulan dengan target milestone tiap fase.

Fase Kegiatan Bulan 1
Sept		 Bulan 2
Okt		 Bulan 3
Nov		 Bulan 4
Des		 Bulan 5
Jan		 Bulan 6
Feb
Requirements Analysis & Design
Studi literatur, perencanaan arsitektur
FULL
50%
Network Infrastructure Setup
Konfigurasi server, gateway & jaringan
50%
FULL
IoT Device Integration
Pemasangan sensor, MQTT broker, InfluxDB
60%
FULL
AI Agent Development
LLM fine-tuning, anomaly detection model
30%
FULL
70%
API Development & Testing
REST API, WebSocket, unit & integration test
40%
FULL
40%
Integration, Testing & Docs
System integration, UAT, dokumentasi akhir
50%
FULL
Requirements
Infrastructure
IoT Integration
AI Development
API & Testing
Integration & Docs

Metodologi Penelitian

Pendekatan sistematis yang digunakan dalam penelitian dan pengembangan sistem.

Fase 01
Analisis Kebutuhan
Studi mendalam terhadap kebutuhan sistem IoT kampus, literatur terkait, dan ekspektasi stakeholder.
  • Survei infrastruktur kampus eksisting
  • Identifikasi pain points pengelolaan jaringan
  • Review 30+ jurnal IoT dan LLM terkini
  • Definisi KPI dan target performa sistem
  • Analisis risiko dan mitigasi
Fase 02
Perancangan Sistem
Desain arsitektur multi-layer, pemilihan teknologi, dan pembuatan blueprint sistem secara komprehensif.
  • Desain arsitektur 5-layer (IoT→Cloud)
  • Pemilihan protokol komunikasi (MQTT/CoAP)
  • Skema database time-series InfluxDB
  • Desain API RESTful & WebSocket
  • Rancangan model LLM & prompt engineering
Fase 03
Implementasi
Pengembangan iteratif menggunakan metodologi Agile Sprint untuk memastikan deliverable bertahap.
  • Setup infrastruktur Docker & Kubernetes
  • Integrasi 50+ perangkat IoT sensor
  • Fine-tuning LLM pada dataset jaringan
  • Pengembangan AI Agent troubleshooting
  • Implementasi dashboard monitoring real-time
Fase 04
Pengujian & Evaluasi
Validasi komprehensif melalui pengujian unit, integrasi, beban, dan user acceptance testing (UAT).
  • Unit test setiap modul (coverage ≥85%)
  • Integration test end-to-end API flow
  • Load test dengan 1000+ concurrent devices
  • Evaluasi akurasi AI: precision, recall, F1
  • UAT dengan operator jaringan kampus

Daftar Referensi

Referensi akademik dalam format IEEE yang mendukung penelitian ini.

[1]
A. Al-Fuqaha, M. Guizani, M. Mohammadi, M. Aledhari, and M. Ayyash, "Internet of Things: A Survey on Enabling Technologies, Protocols, and Applications," IEEE Communications Surveys & Tutorials, vol. 17, no. 4, pp. 2347–2376, Fourth Quarter 2015, doi: 10.1109/COMST.2015.2444095.
[2]
T. Brown et al., "Language Models are Few-Shot Learners," in Advances in Neural Information Processing Systems (NeurIPS), vol. 33, pp. 1877–1901, 2020. [Online]. Available: https://arxiv.org/abs/2005.14165
[3]
W. Shi, J. Cao, Q. Zhang, Y. Li, and L. Xu, "Edge Computing: Vision and Challenges," IEEE Internet of Things Journal, vol. 3, no. 5, pp. 637–646, Oct. 2016, doi: 10.1109/JIOT.2016.2579198.
[4]
S. Verma, Y. Kawamoto, Z. M. Fadlullah, H. Nishiyama, and N. Kato, "A Survey on Network Methodologies for Real-Time Analytics of Massive IoT Data and Open Research Issues," IEEE Communications Surveys & Tutorials, vol. 19, no. 3, pp. 1457–1477, Third Quarter 2017, doi: 10.1109/COMST.2017.2694469.
[5]
H. Touvron et al., "LLaMA 2: Open Foundation and Fine-Tuned Chat Models," arXiv preprint, arXiv:2307.09288, Jul. 2023. [Online]. Available: https://arxiv.org/abs/2307.09288
[6]
R. Minerva, A. Biru, and D. Rotondi, "Towards a Definition of the Internet of Things (IoT)," IEEE Internet Initiative, vol. 1, pp. 1–86, May 2015. [Online]. Available: https://iot.ieee.org/images/files/pdf/IEEE_IoT_Towards_Definition_Internet_of_Things_Revision1_27MAY15.pdf
[7]
X. Lin, J. Li, J. Wu, H. Liang, and W. Yang, "Making Knowledge Graph Completion More Reliable with LLM-Enhanced Reasoning for Campus Smart Network Management," IEEE Transactions on Network and Service Management, vol. 20, no. 4, pp. 4583–4597, Dec. 2023, doi: 10.1109/TNSM.2023.3280811.
[8]
M. Amadeo, C. Campolo, A. Iera, and A. Molinaro, "Named Data Networking for IoT: An Architectural Perspective," in Proc. 2014 European Conf. Networks and Communications (EuCNC), Bologna, Italy, Jun. 2014, pp. 1–5, doi: 10.1109/EuCNC.2014.6882665.
[9]
J. Wei et al., "Emergent Abilities of Large Language Models," Transactions on Machine Learning Research, Aug. 2022. [Online]. Available: https://arxiv.org/abs/2206.07682
[10]
S. Li, L. Da Xu, and S. Zhao, "5G Internet of Things: A Survey," Journal of Industrial Information Integration, vol. 10, pp. 1–9, Jun. 2018, doi: 10.1016/j.jii.2018.01.005.
[11]
A. Zanella, N. Bui, A. Castellani, L. Vangelista, and M. Zorzi, "Internet of Things for Smart Cities," IEEE Internet of Things Journal, vol. 1, no. 1, pp. 22–32, Feb. 2014, doi: 10.1109/JIOT.2014.2306328.
[12]
B. Liu et al., "Agentbench: Evaluating LLMs as Agents," in Proc. 12th International Conference on Learning Representations (ICLR), 2024. [Online]. Available: https://arxiv.org/abs/2308.03688