A team of researchers at Carnegie Mellon University have created an all-in-one sensor which may change the way smart homes are connected in the future. Known as the “Synthetic Sensor,” the prototype can monitor activities in areas of your home and removes the need for consumers to add after-market sensors to products for individual devices to be connected to a network. Inside the device are ten sensors which track and monitor what’s going on in a certain area of your home. The sensors track vibrations, electromagnetic noise, light, and motion and other things to help it to build up a picture of what’s going on, and homeowners can then monitor what’s happening by using their mobile devices. However, the sensor board has some disadvantages. Unlike other smart home products such as Google Home, Synthetic Sensors can monitor what’s happening in your home, but can’t actually be used to control anything. For example, the sensor would be able to inform you whether you’ve left an appliance switched on, but you wouldn’t be able to switch it off remotely unless it’s connected to some other IoT system.
Regardless, this device might help end the problems that many individuals run into when trying to set up smart devices. The devices sometimes struggle to communicate with each other and the only currently viable option to connect many devices from different manufacturers is to fit them with aftermarket sensors afterward, which can be expensive and time-consuming. Synthetic Sensors work immediately and conveniently by simply plugging into a USB wall outlet. Once plugged in, the tiny device uses machine learning technology to assign a unique signature to each object or action, allowing it to distinguish if you are, for example, making coffee or using the washing machine.
According to its creators, the sensors within the device are the key to its success. Machine learning technology was used to train the sensors to recognize the signatures of noises and actions, creating a huge database of noise and actions that the sensors can identify. The researchers have admitted that the universal sensor currently experiences some issues when trying to identify certain noises correctly. This is likely because human environments are constantly changing but this creation should be able to analyze them more efficiently in the near future, its creators claim.