This story was published in the Stacey on IoT newsletter on Friday June 17, 2021.
The quest for more context in the smart home has driven the electronics industry to embrace new forms of sensing — think Ecobee’s decision to put radar into its latest thermostat to better detect people in a home, or Apple’s embrace of ultra-wideband to provide centimeter-level location accuracy for finding items.
Now, two new sensors make clear how much better tomorrow’s sensors could be — and how they will help device makers save on power consumption.
Novelda, a chip firm, introduced a low-power ultra-wideband proximity sensor this week that can detect when people are approaching a device, and then wake up the necessary device components as needed instead of having them run all the time. Novelda estimates this could save between two and 10 years on an LED screen’s life simply by turning it off when no one is around.
This week also saw the introduction of a new form of ultrasonic sensor inside a smart lock made by a Chinese vendor, a debut that could herald the use of modern acoustic sensors in new places in the home. Chinese firm Kaadas is using an ultrasonic sensor that relies on time-of-flight (ToF) sensing from TDK-InvenSense for image recognition of a sorts to help create a smarter smart lock.
When a person walks up to a door, once the ultrasonic sensor determines that they are actually approaching the door it wakes up a processor that checks for image recognition. This is a similar to the architecture found in smartphones that use facial recognition to unlock them; the camera isn’t always on, waiting for a face to appear, as that would waste power.
Instead, a low-power sensor stays on and waits for a face to appear. When it detects one, it wakes up the more robust camera along with a larger processor capable of handling facial recognition. From the user’s perspective, the time between getting their face close to the camera and the unlocking isn’t long, but those two steps help save on battery life.
The Novelda use case noted earlier leverages a similar process. So does the Kaadas lock, but instead of a low-power vision sensor detecting a face, the lock uses an ultrasonic sensor that is capable of millimeter-level accuracy to determine if a person is approaching the door before triggering the higher-power face recognition feature.
Ultrasonic sensors aren’t new. They are already used in multiple settings, especially for things like detecting water levels in tanks or pools. Many of the fountains used for filling water bottles at airports have ultrasonic sensors for tracking when the bottle is in front of the spigot and when it is full. But the new TDK-InvenSense ultrasonic sensor uses ToF sensing and acoustic waves to determine the location and features of a person’s face with much more accuracy.
Instead of just detecting the depth of water or another substance, these sensors can use math to detect the 3-D outline of something using sound waves. It’s similar to how a bat might “see” using sonar. Time-of-flight sensing is familiar to folks working with cameras and lasers as it’s a common method for providing 3-D images for computer vision. The same theory is at work using sound waves.
While today TDK-InvenSense is using the sensor in locks, it has bigger plans. Fabio Pasolini, VP and general manager of the MEMs group at TDK-InvenSense, told me he can’t disclose future use cases, but noted that smarter ultrasonic sensors aren’t only useful when it comes to saving power by waking up a smarter device.
Ultrasonic sensors could provide new functionality to existing products. A sensor could “look” at a glass with liquid in it and determine how much liquid is inside, for example, then calculate how many more ounces of the liquid would be needed to fill the glass with 12 ounces of water.
That level of accuracy is new, and would be especially interesting in refrigerator water dispensers or even sinks. So I’m going to stay tuned for whatever new use cases come out for the TDK-InvenSense sensor and the Novelda one, and generally keep an eye out for a future of devices with more sensing capabilities designed to save power, provide more context, and generally make my gadgets a little bit more compelling.