AI-Powered Smart Lighting Control from Sensor Data

Motion sensors are a cheap way to control lighting, but they are blind: lights turn on with delay, turn off when someone is still in the room, and ignore daylight. The result — 30% of energy is wasted. We, a team with over 5 years of experience in ML building automation (more than 50 completed projects), solve this problem differently: we train a model to predict occupancy and smoothly adjust brightness to real conditions. The system pays for itself in 2–4 years, and staff comfort increases — no harsh switching or flicker.

In one project for an office building of 2500 m², we implemented a system based on Edge ML. We installed 12 Raspberry Pi 4 units connected to a DALI network. The Random Forest model was trained on 3 months of historical data and achieved 94% occupancy accuracy. Energy consumption dropped by 45%.

How AI Solves Blind Lighting

Classic PIR sensors provide a binary signal: "motion/no motion." They cannot distinguish a person from a cat, do not remember history, and do not know that at 3:00 PM the meeting room is usually occupied. Our approach — an ML model on time series — collects data from sensors (PIR, ultrasound, CO2, lux) and builds a probabilistic forecast. The Random Forest model yields 95% occupancy accuracy — 1.5 times more accurate than typical PIR (70%). Daylight harvesting based on autoregression reduces brightness when natural light reaches 500 lux — an extra 15% savings.

What the System Does

• Turns off lighting when no people are present after N minutes (adaptive timeout per zone, considering return probability).
• Reduces brightness when daylight is sufficient — the model predicts illuminance one hour ahead.
• Pre-lights before people arrive (based on calendar/patterns).
• Emergency lighting when motion is detected in dark periods — with smooth ramp-up to 20%.
• Automatic calibration: the model retrains every week on new data, adapting to seasonality.

Why Choose ML Over Simple Timers?

Timers are inflexible: on Friday evening the office is empty, but lights stay on until 11:00 PM. An ML model analyzes occupancy over the last 8 weeks and predicts empty zones with probability 0.97 — lights turn off 40 minutes earlier. The "last person leaves" scenario saves up to 8% of total consumption. Plus, the model suppresses false triggers from drafts and animals — detection accuracy does not drop. The ML model cuts energy consumption by 30–50%, which is 2–3 times more than timers (10–20%).

Impact Metrics

Comparison of Lighting Control Approaches

What's Included in the Work

1. Site survey: audit of current lighting system, zone diagrams, illuminance measurements, historical sensor data collection.
2. ML solution design: model selection (Random Forest, LSTM, Prophet), hyperparameter tuning, simulation on synthetic data.
3. Installation and integration: edge controller setup, connection via DALI/Modbus, model deployment through ONNX Runtime. Integration with BMS via BACnet, Modbus, KNX is possible.
4. Calibration and A/B test: 2–3 weeks of parallel operation, comparison with existing logic.
5. Staff training and documentation: handover of dashboard (Grafana), API for BMS integration.

Ready to estimate savings for your building? Contact us for a preliminary calculation.

Timeline: 4–8 weeks

Leave a request for a savings calculation — we will assess your facility in 2 days and calculate payback. We guarantee at least 30% reduction in energy consumption.

DALI — standard IEC 62386 for digital lighting control.

Edge AI and Optimization: How to Deploy Models Without Cloud?

Imagine: your face recognition model has 4 seconds latency on Jetson Orin, the battery runs out in an hour, and the model crashes with OOM. We are a team of Edge AI engineers with 5+ years in production — we have optimized over 150 models for edge devices. Without profiling and proper choice of quantization or distillation, the project is doomed. The gap between research code and edge deployment is a separate engineering discipline; we help you master it in 2–16 weeks turnkey. Edge AI and model optimization services are not just export, but systematic work with hardware.

Why Simply Exporting a Model Doesn't Work?

A PyTorch model with float32 and batch_size=32 is not ready for edge. Typical problems:

• ResNet-50 in fp32 occupies 98 MB, inference on Cortex-A78 — 380 ms. After INT8 quantization via torch.ao.quantization — 24 MB, 95 ms. Export to ONNX + TensorRT on Jetson — 28 ms.
• YOLOv8m on Raspberry Pi 5 in fp32 — 2.8 fps. TFLite INT8 — 9.4 fps. With XNNPACK delegate — 14 fps (1.5× faster than pure INT8).
• Transformer encoder on mobile CPU: MobileBERT in fp16 via CoreML on iPhone 15 — 18 ms/inference. distilbert-base-uncased in ONNX — 42 ms.

The problem is not choosing "quantize or not" — the right path is determined by the device, task, and acceptable metric degradation. We offer an assessment of your project: within 24 hours we will tell you how feasible it is to speed up the model.

How to Choose Quantization Method for Your Task?

PTQ (Post-Training Quantization) — a quick path. Take a trained model, run a calibration dataset (200–1000 samples), get INT8 or INT4 weights. Tools: torch.ao.quantization, ONNX Runtime quantization tool, bitsandbytes. Accuracy degradation: 0.5–2% on classification. Red zone — small object detection and segmentation, where PTQ gives -4–8% mAP.

QAT (Quantization-Aware Training) — training with simulated quantization noise. More expensive (retraining), but degradation 0.1–0.5%. Justified when PTQ is unacceptable. In PyTorch — torch.ao.quantization.prepare_qat().

GPTQ / AWQ — for LLMs. AWQ better preserves quality at 4-bit quantization. llm-compressor from Neural Magic or autoawq are the main libraries.

Potential savings from choosing the right method can be substantial — for example, reducing cloud inference costs by up to 70% when deploying to edge. Project cost is calculated individually based on model complexity and target platform.

When to Use Pruning vs Distillation?

Structural pruning removes channels or layers. torch.nn.utils.prune — basic tool. For transformers — attention head pruning (LTP, movement pruning). Result: ResNet-50 after removing 40% of channels with fine-tuning — -35% size, -28% latency, -1.2% top-1 accuracy.

Knowledge distillation — train a small student to mimic a large teacher. Classic via KLDivLoss on soft labels. Feature distillation on intermediate layers is more effective. Hugging Face DistilBERT: 66M vs 110M parameters, -40% latency, -3% on GLUE. This is a model compression technique.

Combined approach: distillation → pruning → QAT. Gives maximum effect on limited hardware. We recorded a case where a client achieved 70% reduction in cloud compute spend after moving to edge with this pipeline.

Target Platforms and Tools

TensorRT — the main tool for NVIDIA edge. TRT builds a graph with operator fusion, selects optimal kernels. On Jetson AGX Orin YOLOv8m in TRT INT8 gives 78 fps vs 22 fps in fp16 PyTorch — 3.5× improvement.

Practical Case: How We Detected Defects on a Production Line (Our Client)

Task: real-time scratch detection on metal, 30 fps, camera to Jetson Xavier NX (16GB). Original model YOLOv8l mAP50 0.91, server inference 28 ms, on Jetson in fp16 — 110 ms (9 fps). Not suitable.

Optimization steps we performed for our client:

1. Switch to YOLOv8m — mAP50 0.887 (-2.3%), 68 ms
2. Export to TensorRT FP16 via yolo export format=engine half=True — 31 ms (32 fps)
3. INT8 calibration on 500 frames — 22 ms (45 fps), mAP50 0.879

Result: 3.5% degradation at 5× speedup. Client received engine and documentation. We guarantee metric will not drop below agreed threshold — specified in contract.

What is Included in the Work?

• Report on model profiling on target device (layer latency, bottlenecks)
• Selection and justification of optimization methods (quantization / pruning / distillation)
• Optimized model (TensorRT engine / TFLite / CoreML / ONNX)
• Configs for reproducibility (scripts, Docker image, instructions)
• Testing on real device (at least 10,000 inferences)
• Training of your team (2 hours online)
• 1 month support after delivery

How to Order Model Optimization

1. Submit a request on the website or contact us in any convenient way.
2. We perform free profiling of your model on the target device within 24 hours.
3. We prepare an optimization plan with trade-off estimates (speed vs quality).
4. You approve the plan — we start work.
5. After completion, we deliver the optimized model, configs, and documentation.
6. We train your team and provide monthly support.

Timeline: optimization of an existing model — 2–4 weeks. Development from scratch for edge — 6–16 weeks.

Get a consultation — we will evaluate your model for free and offer a plan within 24 hours. Order free profiling now. For complex projects, contact our engineering team to discuss custom optimisation strategies.