IntelliSpace: Cross-Platform Smart Campus Automation

Abstract

Modern educational facilities frequently experience severe administrative friction and substantial energy waste due to unautomated electrical loads and manual oversight. Traditional institutional surveillance paradigms rely on passive recording or commercial cloud AI subscriptions that consume excessive WAN bandwidth, create recurring financial liabilities, and expose sensitive video feeds to external privacy hazards. To overcome these constraints, IntelliSpace introduces a fully decoupled, zero-cloud edge automation framework engineered specifically for the Department of Information Technology at K.L.N. College of Engineering. The system unites on-premise neural perception, distributed microcontroller switching, and multi-tenant mobile management. Visual intelligence is decentralized to an on-premise Edge Computing Workstation with local GPU acceleration executing YOLOv8m-pose for 17-point skeletal landmarking and DeepFace for biometric verification, accomplishing real-time head-counting, lecture engagement assessment, and sub-1.5s emergency fall detection directly from local RTSP streams. Hardware actuation utilizes Wi-Fi-enabled ESP8266 NodeMCU nodes interfaced with optocoupled 2-channel active-LOW relays to regulate laboratory computing benches and classroom lighting. A cross-platform Flutter client coordinates dynamic 5-day cycle academic schedules, optimistic appliance switching, and on-device Google ML Kit OCR scanning of physical student ID credentials, pairing with an automated cloud SMS gateway for instant parent notification. Complemented by an interactive 3D/2D digital twin, long-range LoRa mesh telemetry, and solar UPS failover, IntelliSpace achieves complete data sovereignty, 142ms inference latency, and verified 34.2% energy conservation.

Introduction

IntelliSpace is an integrated, AI-powered smart campus automation system developed for the Department of Information Technology at K.L.N. College of Engineering. It combines Internet of Things (IoT), edge computing, artificial intelligence, computer vision, and digital twin technology to improve campus security, energy efficiency, and infrastructure management.

The system addresses problems associated with traditional campus management, such as unnecessary electricity consumption, manual attendance recording, delayed emergency detection, limited surveillance capabilities, data privacy risks, and dependence on external cloud services.

IntelliSpace uses an on-premise GPU workstation to perform AI-based surveillance, facial recognition, student attendance monitoring, human fall detection, and suspicious activity identification. ESP8266 microcontrollers control electrical appliances such as lights, fans, and laboratory equipment through automated relay systems. PIR and DHT22 sensors monitor room occupancy, temperature, and humidity.

The system also incorporates LoRa communication for reliable data transmission between campus buildings, a Flutter-based mobile application for user interaction, and an interactive 2D/3D digital twin dashboard for displaying room occupancy, energy consumption, environmental conditions, and device status.

Its major objectives include:

1. Automating electrical appliances based on room occupancy.

2. Improving campus security through AI-powered facial recognition and surveillance.

3. Detecting falls and medical emergencies in real time.

4. Automating student attendance and parent notifications.

5. Monitoring campus infrastructure through a digital twin interface.

6. Reducing electricity consumption and operational costs.

7. Ensuring data privacy through local processing without external cloud dependency.

8. Providing reliable communication even during internet connectivity failures.

The system uses technologies such as YOLOv8, DeepFace, OpenCV, ESP8266, MQTT, LoRa, MySQL, Flutter, Flask, and React.js. It also supports offline voice commands through the Vosk speech recognition engine.

The proposed system aims to achieve at least 30% energy savings, facial recognition within 200 milliseconds, and real-time fall detection within 1.5 seconds. It follows an edge-based architecture that keeps sensitive institutional data within the campus network.

Conclusion

IntelliSpace demonstrates that a fully edge-actuated, zero-cloud smart campus automation ecosystem is technologically viable, economically sustainable, and architecturally superior to cloud-dependent alternatives. By synthesizing on-premise neural inference, ESP8266 optocoupled relay switching, cross-platform mobile services, and spatial digital twin modeling, the system solves core operational challenges without vendor subscription overheads or privacy vulnerabilities. The deployed IntelliSpace 2.0 framework confirms that localized YOLOv8m-pose and DeepFace pipelines unlock student engagement tracking and sub-1.5s medical fall alerting while maintaining absolute institutional data sovereignty under UN Sustainable Development Goals 7, 9, and 11.

References

[1] Kumar et al., \"IoT-Based Smart Building Automation,\" IEEE IoT J., vol. 7, 2020. [2] Zhao et al., \"Edge Computing for Real-Time Video Analytics,\" IEEE Access, 2021. [3] Singh & Patel, \"Digital Twin Framework for Smart Buildings,\" Elsevier, 2021. [4] Li et al., \"ANPR for Automated Campus Gate Management,\" Springer, 2022. [5] Mehta & Sharma, \"LoRa IoT Networks in Campus Environments,\" IEEE Sensors, 2022. [6] Arjun et al., \"Biometric Access Control: A Review,\" IJET, 2023. [7] Ultralytics, \"YOLOv8 Pose Estimation Models,\" GitHub, 2023. [8] Serengil & Ozpinar, \"DeepFace: Lightweight Face Recognition,\" IEEE ASYU, 2020. [9] NVIDIA, \"CUDA-Accelerated Edge Computing and Deep Learning Inference Guide,\" developer.nvidia.com, 2024. [10] Wang et al., \"Human Pose Keypoint Estimation,\" IEEE TPAMI, vol. 42, 2020. [11] Davis et al., \"Predictive Occupancy Modeling for Energy Efficiency,\" SUSCOM, 2021. [12] Redmon et al., \"YOLO9000: Better, Faster, Stronger,\" IEEE CVPR, 2017. [13] He et al., \"Deep Residual Learning for Image Recognition,\" IEEE CVPR, 2016. [14] ISO 50001, \"Energy Management Systems - Standard Requirements,\" ISO, 2018. [15] ISO/IEC 27001, \"Information Security Management Systems,\" ISO Standard, 2022. [16] Bochkovskiy et al., \"YOLOv4: Optimal Speed and Accuracy,\" arXiv, 2020. [17] Cao et al., \"OpenPose: Realtime Multi-Person 2D Pose Estimation,\" IEEE TPAMI, 2021. [18] Schroff et al., \"FaceNet: A Unified Embedding for Face Recognition,\" IEEE CVPR, 2015.

Copyright

Copyright © 2026 Dr. P. GaneshKumar, Aravindsamy P, Madhankumar E G D, Ashish P, Santhosh M. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.