The "ambient, socially aware devices" use Soli Radars, machine learning, and deep learning to interpret our actions.

Google in its latest effort is trying to blend the Radar and motion-sensing technology in the corners and places of your integrated home environment.

In a recent YouTube video, Google's ATAP group that is behind the wearable technology and the Soli Radar in the Pixel 4, showed off its new research project on non-verbal interactions with the radar system. The researchers show how they use a mesh network of Soli Radars, which can be built into things like Google Nest, to keep track of people as they move around and near things.

Google explained that it is aiming for devices to understand what it refers to as an environment's "social context" through Soli-recognized nonverbal behaviors such as "approach and leaves," "turning toward/away," and glance. This includes the ability to recognize "when someone approaches or enters its personal space."

The "ambient, socially aware devices" use Soli Radars, machine learning, and deep learning to interpret our actions. A hand wave, a head turn, distance from devices, speed past them, and even body orientation tell the system something about our intent.

Google ATAP Head of Design Leonardo Giusti says "People inspire us. Humans intuitively understand one another without saying a word. We pick up on subtle social cues, subtle gestures that we understand and react to.”

When it comes to practical applications, the company demonstrated a square-shaped smart display mounted on a wall near a home's front door. When no one is present, the temperature is the primary display, with the forecast being noted in smaller text below the temperature display. The forecast takes up the entire screen as the plane approaches.

Additionally, a square display that tells you what song is playing and when new alerts are received is another example of this. The entire message is only visible if you keep your eyes glued to the screen. Once you've stopped looking, the full music user interface will be revealed.

A tablet and a kitchen display are the other two examples. When you walk up to the device, it automatically answers an incoming video call. If you walk away from your tablet while a video is playing, the video will be paused.

Because radar does not "see" anything and does not collect images from the environment, it is preferred over optical sensors for intent detection. Radio waves create a motion mesh that is later programmed in the software to detect and interpret.