Researchers from Rensselaer built smart lighting systems using distributed color sensors

Share Article

Signal Analysis and Machine Perception Laboratory, Rensselaer Polytechnic Institute presents their low-cost and privacy-preserving occupancy sensing technique for monitoring indoor human activities

In the U.S., about 11% of total electricity consumption is used for lighting

Past News Releases


Researchers from Rensselaer Polytechnic Institute presented a smart lighting system which can intelligently optimize the lighting condition based on human activities at International Conference on Pattern Recognition in Stockholm.

Lighting systems are installed in every home and office. They are used by everyone, and have been an essential part of modern lives. According to the U.S. Energy Information Administration, about 11% of total U.S. electricity consumption was used for lighting in the year of 2014. The percentage is 15% for residential and commercial sectors. The task of smart lighting is to improve our current lighting systems, to reduce energy consumption, and to enhance human comfort, well-being, and productivity.

A great effort in smart lighting is occupancy sensing, which is to use particular sensors to monitor the use of the space, such that light is only produced when and where it is needed. Existing occupancy sensing solutions have all kinds of drawbacks. For example, camera-based occupancy sensing is not only expensive, but also raises concerns of privacy and information security.

The Signal Analysis and Machine Perception Laboratory at Rensselaer Polytechnic Institute invented a novel solution, named COlor-Sensor-Based Occupancy Sensing, or simply COSBOS. The sensors used by them are very cheap photodiode-based color sensors that do not capture any images. Thus their sensing system is very low-cost, and unlike using cameras, it protects the privacy of users. Their system is based on two key ideas. The first idea is to modulate a small signal to the light emitted from the LED lighting fixtures. This signal is very small, thus cannot be observed by human, but can be measured by color sensors. The measured signal from the sensors is used to recover a "light transport model" - a mathematical model to describe how photons transport in the space. The second idea is to use machine learning techniques to predict where blockage and reflection happen in the space, and reconstruct a "heatmap" showing where human activities are in the room.

“With this occupancy sensing technique, we can monitor the use of an indoor space, and design intelligent illumination control strategies to only deliver light at the right place and the right time, which will greatly reduce energy consumption,” said Dr. Quan Wang, the first author of this work.

“If this technique becomes commercialized in the future, it would potentially change the way people interact with the space they live in,” according to a post from Technology Org on December 2, 2015:

Additional information can be obtained from Dr. Quan Wang, Rensselaer Polytechnic Institute, or from their project website:

Reference: Wang et al., Occupancy distribution estimation for smart light delivery with perturbation-modulated light sensing. Journal of Solid State Lighting 2014 1:17.

Share article on social media or email:

View article via:

Pdf Print

Contact Author

Shulin Han
Energy Observer Magazine
+86 18515236786
Email >
Visit website