Digital lighting infrastructure provides an excellent basis for generating data. To get to the treasure, however, you first have to dig.

In his book “Megatrends”, John Naisbitt, the American author and speaker on future studies, wrote: “We are drowning in information but starved for knowledge.” That was back in 1982 and the sentence has been often quoted since. Never has it been more relevant than today. The digitalization of life is leading to an ever rising flood of data. But the data is useless unless it is suitably processed, analyzed and understood.

The magic word that separates the data wheat from the data chaff is “data mining.” It implies statistical methods and algorithms used to track down patterns and relationships within large volumes of data. “Data mining can uncover hidden information but cannot say anything about its value. Companies therefore have to learn to understand their data and interpret it correctly,” says Louis Trebaol. The American heads the data analytics team at OSRAM.

He and his team are investigating new applications and business opportunities resulting from the data arising from OSRAM’s products and systems. These are producing more and more data, such as direct lighting data generated in the course of the digitalization of light. There is also an increasing amount of data from the growing number of sensors being enabled through the lighting infrastructure. This infrastructure is omnipresent, powered and perfectly positioned to service a wide range of sensors – whether as streetlights or ceiling luminaires – to cover large areas.

“Up to now, sensors have provided us above all with information on how we can better control lighting systems. For example, we can use presence detection in offices to identify typical attendance patterns and adjust the light to actual demand,” says Trebaol. “In the future, we intend to use the data primarily to make people and machines more efficient.” For instance with human centric lighting solutions in which the light can be adapted according to the time of day, the amount of available daylight and the biorhythms of individual employees.


Machine failing and learning

“We consider data analytics to be more than just searching for existing patterns; it’s also about developing new assumptions,” added Trebaol. “It is a form of machine learning.” His favorite example is the cultivation of plants in a controlled environment, known as “smart farming.” In the search for the right light for optimum plant growth, our multi-channel LED luminaires and connected sensors are supplying a wealth of data. “From this data we can derive successful growth patterns. But we can also test new lighting scenarios and develop new hypotheses from the results. The larger the volume of data, the greater the demand for artificial intelligence which can learn from the data and make independent decisions about controlling the light.”

Linked to this principle of trial and error is a new way of thinking in the company. “Our approach needs to be focused less on products and more on solutions. The product business is dying. It is moving more toward services that are activated by hardware.” OSRAM is therefore embarking on a new path with its “Lightelligence” IoT platform and offering its customers data services around its products. “We are starting with lighting. But we must be prepared to follow customers wherever their demands take us.” Data analytics can help here to flag new demands and opportunities in the first place. Coming from a startup, Trebaol knows that this sometimes means trying a number of different approaches. “We have the capability to create all kinds of solutions. The real question is this: Is the solution really important to our customers, and are they prepared to pay for it?”

 

The mix matters

To find this out, Trebaol has set up a team of diverse experts. “We have an interesting mix of academic and professional backgrounds in our team – from physicists and computer scientists to financial people and industrial engineers. What unites them is their impressive ability to model available data with the aid of artificial intelligence. And they can relate to technology and to business so they can answer two questions: Can we do it? And should we do it?”

And what drives him personally? “I am really passionate about smart farming. If in a few years’ time we can say that OSRAM played a significant role in automating and thereby improving the food supply for mankind I would be proud.”