Artificial intelligence versus stand-by
Office equipment is often left on overnight, consuming a lot of power. That doesn't have to be the case. The Lucerne University of Applied Sciences and Arts and an industry partner have developed a component that uses artificial intelligence to know when to shut down devices in a controlled manner.
If electrical devices are not switched off at night, they often consume more energy on standby than during the day when in use. This is particularly true for office equipment such as printers or coffee machines, which often nobody is responsible for switching off, especially in large offices. Together with a team from the iHomeLab at the Lucerne University of Applied Sciences and Arts (HSLU), Schurter AG has therefore developed a component for office and other equipment that is intended to remedy this unnecessary energy consumption with the help of artificial intelligence - up to 30 percent can be saved with its help.
Artificial intelligence detects operation without benefit
This compact component, which manufacturers can easily integrate into their devices or retrofit, uses power consumption to determine when devices are actually needed, according to the HSLU media release. "The system collects data over a period of time and recognizes patterns," explained Andrew Paice, director of the iHomeLab. "In this way, it learns during which times a device is in so-called operation without benefit and therefore can and should be switched off completely." Once the pattern is recognized, the component shuts down the corresponding device at night in a controlled manner. However, it can be switched on again manually at any time - unlike, for example, if a timer were used. In addition, the artificial intelligence quickly recognizes when the behavior of employees changes at short notice, for example, because a project deadline is looming and the working days are therefore longer for a while. The time for shutdown is automatically adjusted accordingly. The component could also have reacted to the sudden absence of employees due to Corona.
The iHomeLab team's task was to develop algorithms that could be used to explore and discover "operation without utility." "In addition, we had to make sure that the algorithms worked with the right data. For example, in the case of the water dispenser, power consumption alone is not an indication because cooling requires power even if no one is drawing water; however, the algorithm must only respond to power consumption due to water draw," Paice explains.
Computing power thanks to cloud
To recognize the patterns in the behavior of, for example, an office crew, the data must be sent to a cloud - without drawing conclusions about the individual employees and their personal data. According to HSLU, the artificial intelligence algorithms reside in the cloud to provide the computing power needed to recognize the patterns. This allows the devices to be connected to the Internet of Things (IoT), he said. The IoT connection is particularly useful for manufacturers who want to learn more about the typical use of their products in order to improve them. In the case of a water dispenser, when does water need to be refilled? Is a maintenance service due? Is the water dispenser in the right place - where most employees can use it? Thanks to these algorithms and the space-saving design, the Smart Connector can be used in virtually any device, according to the HSLU.
Source: Lucerne University of Applied Sciences and Arts
