Beginner IoT Development · Beginner
Sensor Fusion Basics for Occupancy Experiments
Blend PIR and CO2 signals without overfitting to one office layout.
Overview
Design lightweight fusion rules, cross-validate against synthetic ground truth, and document failure cases when HVAC cycles distort readings.
Included focus areas
- Calibration drift notebooks
- Synthetic occupancy generators
- Cross-validation splits by room geometry
- Bias-variance discussion grounded in real traces
- Ethical placement guidelines
- Visualization palettes readable under dark UI themes
- Export formats compatible with BI stubs
Outcomes you can evidence
- Ship fusion notebook with peer-reviewed assumptions
- Present failure cases with sensor-specific caveats
- Demonstrate rollback to single-sensor fallback
FAQ
Dataset size?
Classroom-scale only—no enterprise warehouse dumps.
Python depth?
Pandas-light; heavy ML pipelines not covered.
Office hours?
Biweekly; priority queue by enrollment date.
Participant notes
“CO2 vs PIR conflict lab mirrored our office ghost readings—finally have language for HVAC interference.”
“Notebook exports slotted into our BI stub faster than expected.”