In the next five years, billions of hands-free, always-on sensing devices that run on battery will assist us in our daily lives, at home and at work. Some of these devices will be always-listening for a voice command to play music or to control the temperature and lighting in our homes, while others will be always-feeling the vibrations of industrial equipment for early detection of equipment malfunction and predictive maintenance. These are just some of the always-on sensing applications that are already in use today ― so imagine all the new applications that are yet to come.
The small size and portability of many always-on sensing devices makes them very attractive for deployment throughout the home or the factory floor. The small batteries that keep these devices going, however, present a fundamental design problem because these devices are continuously processing high-bandwidth data and are prone to power- and data-overload issues that drain the battery. Requiring consumers to frequently recharge or replace batteries is a deal-breaker for many applications. Who wants to recharge their smart earbuds every few hours or regularly replace the batteries on multiple condition-monitoring sensors deployed throughout a factory?
The problem is that the traditional “digitize-first” system architecture for always-on sensing devices digitizes and analyzes all the incoming sensor data, expending precious power and data resources on data that is mostly irrelevant and will ultimately be thrown away. While lower-power sensors and other digital processing components can incrementally help to reduce system power, a more efficient system-level approach that minimizes the amount of data handled by the system can achieve unprecedented power savings.
Analyze-First: A New System Architecture for Always-on Sensing
During Sensors Expo 2019, we will introduce attendees to a new power- and data-efficient system architecture: an “analyze-first” edge architecture that leverages Aspinity’s RAMP (Reconfigurable Analog Modular Processor) technology to analyze the vast amount of data gathered by sensors – before the data is digitized. By directly analyzing the raw analog sensor data for what’s important, our “analyze-first” architecture reduces the volume of sensor data that is processed through higher-power system components (e.g., digital processors and ADCs) by up to 100x and reduces always-on system power by 10x. This architectural shift from “digitize-first” to “analyze-first” translates into significant end-user benefits for the next generation of portable always-on sensing devices, enabling, for example, smart earbuds that last a full day on a single charge or a voice-first remote control that can last a year on a single battery.
For additional information about using an “analyze-first” architecture for always-on sensing, please attend the Sensors Expo Pre-Conference Symposia on Tuesday, June 25, where Aspinity CTO/Founder Brandon Rumberg, Ph.D. will join The Future of MEMS & Sensor Engineering panel (Pre-Conference Symposium: Using MEMS and Sensor Technology to Improve the Future), and CEO/Founder Tom Doyle will present Boosting Intelligence at the Edge for Power- and Data-Efficient Industrial Sensing (Pre-Conference Symposium: Innovation at the Edge). Brandon Rumberg will also present Overcoming the Challenges of Voice-First for IoT on Thursday, June 27 during Sensors Expo. To meet with Aspinity at Sensors Expo, please contact us: [email protected].
About the authors
Tom Doyle is the founder and CEO and Brandon Rumberg is the co-founder and CTO of Aspinity