Multi-channel low-cost light spectrum measurement using a multilayer perceptron

• Published in: Energy and buildings Vol. 199; pp. 579 - 587
• Main Authors: Botero-Valencia, J.-S., Valencia-Aguirre, J., Durmus, D., Davis, W.
• Format: Journal Article
• Language: English
• Published: Lausanne Elsevier B.V 15-09-2019; Elsevier BV
• Subjects: Artificial neural networks; Bluetooth; Electric power distribution; Internet of Things; Light; Light sources; Low cost; Multi spectral sensor; Multilayer perceptron; Multilayer perceptrons; Neural networks; Spectral power distribution; Spectrometers; Spectrum measurement; Visible spectrum; Wireless communications; Internet of things; Multilayer perceptron; Spectral power distribution; Spectrum measurement; Multi spectral sensor
• ISSN: 0378-7788; 1872-6178
• DOI: 10.1016/j.enbuild.2019.07.026

•Light is one of the most important elements for residential and work spaces.•Measuring light source spectrum with spectrometers is expensive.•A low-cost spectrometer was developed using an Artificial Neural Network.•The reconstructed SPD has an error lower than 2%.•The device has wireless communication (Bluetooth and Wi-Fi).
Light is one of the most important elements for residential and work spaces, which affects visual performance, comfort, productivity and well-being. The measures that quantify the characteristics of a light source are derived directly from the spectral power distribution (SPD). In addition the SPD is an important factor influencing the quality of a light source. However, measuring light source spectrum with traditional spectrometers is expensive, difficult to adapt to normal spaces, and hard to integrate with other systems. To address these challenges, a low-cost spectrometer was developed using an Artificial Neural Network, with a resolution of 5 nm in the visible spectrum. The reconstructed SPD has an error lower than 2% and allows the derivation of measurements to characterize the colour quality of light sources. Additionally, the device has wireless communication (Bluetooth and Wi-Fi) in real time, which allows integration into lighting control applications and other Internet of things (IoT) applications.