Infineon Applied sciences AG is introducing its battery-powered Good Alarm System (SAS). The expertise platform achieves excessive accuracy and really low-power operation utilizing sensor fusion based mostly on synthetic intelligence/machine studying (AI/ML). This expertise mixed with low-power wake-on acoustic occasion detection offers exceptional efficiency. The compact design exceeds the detection accuracy of acoustic-only alarm programs used at the moment in sensible buildings and houses, and different IoT functions whereas reaching equal or higher battery life in comparison with much less subtle options.
“We’re excited to allow a singular and differentiated method to bringing AI/ML capabilities to cost-sensitive, battery-powered house safety sensor programs, with out sacrificing battery life,” mentioned Laurent Remont, Vice President of IoT and Sensor Options at Infineon’s Energy & Sensor Methods Division. “Present house safety options are unreliable for detecting occasions resembling glass break. Our new resolution combines quite a lot of best-in-class applied sciences to create an alarm system that’s sensible, dependable and energy environment friendly. We sit up for bringing extra progressive options into the house safety market.”
The answer incorporates Infineon’s excessive signal-to-noise ratio (SNR) analog XENSIV MEMS microphone IM73A135V01 , XENSIV digital stress sensor DPS310 and PSoC 62 microcontroller. Infineon additionally offers a sensor fusion software program algorithm based mostly on exactly educated AI/ML that mixes acoustic and stress sensor knowledge to precisely differentiate between sharp sounds inside a house and distinctive audio/stress occasions. These occasions might be created when glass is damaged, or a home alarm is triggered on account of a smoke alarm, a carbon monoxide alarm or an intrusion is detected by a door or window.
The AI/ML sensor fusion algorithm can also be able to eliminating many different background sounds or background stress occasions that may generate false positives as a result of similarities to alarm programs.
