SmartLight AI/ML - Pixelegg
SmartLight AI, Vision-Driven Energy Management Reduction light time: ↓55% and Centrally monitored 100% AI-powered smart lighting that combines schedules, real-time occupancy, and IoT control to cut wasted energy in campuses and offices. Key Results Reduction in wasted lighting time: ~40–55% False “lights off during use” incidents: ↓ ~80% (vs. basic motion sensors) Rooms centrally monitored: 100% of connected classrooms and offices Decision signals combined: 2 (time schedule + AI occupancy) for every switching action
Client
Pixelegg GmbH
Industries
Education, Corporate Offices, Smart Buildings, Hospitality
Services
AI/ML (Computer Vision), IoT Edge Engineering, Cloud Backend, Web Dashboard UX
Project Scope
Vision-Driven Smart Lighting System Implementation
Challenge
Institutions needed a precise, AI-driven solution that could determine when a room should be lit based on schedules and whether it was actually occupied using real-time vision, while keeping everything manageable from a single, unified dashboard. Traditional systems fall short: manual switches leave lights burning in empty rooms when someone forgets, timer-based controls can’t adapt to extra classes or last-minute meetings, and basic motion sensors often misread still classrooms, abruptly turning lights off mid-session.
Compounding these issues, most campuses and corporate buildings lack any centralized visibility. Facility teams can’t see which rooms are currently lit, can’t override or adjust behavior remotely, and have no reliable way to analyze usage trends across buildings. The result is wasted energy, disrupted sessions, and no data to improve operations, driving the need for a smarter, integrated approach.
Solution
SmartLight AI: Intelligent Lighting Control for Modern Campuses
NITSAN built SmartLight AI as a unified platform combining IoT edge devices, cloud-based scheduling, and computer-vision occupancy detection for precise, automated lighting control across buildings.
IoT Scheduling with Remote Control
A web dashboard allows facility teams to configure daily/weekly time slots for each room, store schedules in the cloud, and push them to edge devices. Administrators can override schedules instantly for exams, events, or emergencies, no physical access needed.
AI-Powered Occupancy Detection
During active time slots, ceiling- or wall-mounted cameras capture room images. A CNN model (MobileNetV2 / YOLO-style) analyzes presence and outputs a probability P(O). Lights are switched on only when the schedule is active and P(O) exceeds a defined threshold (e.g., 0.8).
Centralized Monitoring & Logging
The dashboard provides real-time visibility into all rooms, their on/off state, active schedule, and last occupancy decision. Historical logs track all switch events, detections, and overrides for energy reporting and space-usage analysis.
Multi-Room, Multi-Building Support
A single cloud backend manages multiple buildings and classrooms, each with independent schedules and devices, making the system scalable for campuses and large organizations.
AI Architecture Flow
AI Technical
Stack
Node.js
REST & MQTT APIs
HTML
CSS
JavaScript
0%
Lights operate only when rooms are scheduled and occupied, outperforming manual, timer-based, or motion-sensor setups.
0%
Fewer “lights-off mid-session” incidents thanks to robust, vision-based occupancy detection.
0x
Facilities teams can schedule, monitor, and override dozens of rooms from a single dashboard with minimal effort.
0%
Extend control to ACs, projectors, and other smart building systems using the same IoT + AI framework.
AI Architecture Flow
Client's
feedback
“SmartLight AI gave us real control over energy use. Classrooms now light themselves only when schedules and real occupancy align, and our facilities team finally has a live view of what’s happening across campus.”
Melanie Reinhardt
Operations VP Pixelegg GmbH
Your Next Project is a Conversation Away!
