Home Electronics News New Products Home Security Sensor Systems With AI/ML Capabilities

Home Security Sensor Systems With AI/ML Capabilities

Home Security Sensor Systems With AI/ML Capabilities

Infineon Technologies has released the company’s first Smart Alarm System (SAS) that is powered by batteries. The system incorporates Infineon’s XENSIV MEMS microphone IM73A135V01. The SAS utilises artificial intelligence (AI) and Machine Learning (ML) to achieve high accuracy while consuming minimal power operation using sensor fusion. The system can be used for a home security system as an intrusion detection system and it is capable of detecting low levels of sounds with greater accuracy which was not possible of less sophisticated solutions. The solution is capable of differentiating different sounds such as the breaking of glass or an alarm triggered due to fire.

“We are excited to enable a unique and differentiated approach to bring AI/ML capabilities to cost-sensitive, battery-powered home security sensor systems, without sacrificing battery life,” said Laurent Remont, Vice President of IoT and Sensor Solutions at Infineon’s Power & Sensor Systems Division. “Current home security solutions are unreliable for detecting events such as glass break. Our new solution combines a number of best-in-class technologies to create an alarm system that is smart, reliable and power efficient. We look forward to bringing more innovative solutions into the home security market.”

The SAS system incorporates a XENSIV digital pressure sensor DPS310 along with a high signal-to-noise- (SNR) analog XENSIV MEMS microphone IM73A135V01 and PSoC 62 microcontroller. The battery-powered SAS device uses a sensor fusion software algorithm which has been precisely trained using AI/ML that combines acoustic and pressure sensor data to accurately differentiate between sounds of different objects and events. It is capable of differentiating the sharp sounds inside a home and distinctive audio/pressure events. The SAS device can accurately identify events such as when a glass is broken, or a house alarm is triggered. The trained model is capable of eliminating background noises or background pressure events that can generate false positives due to the similarities to alarm systems.