STMicroelectronics’ first AI-enhanced smart accelerometer: improving performance and energy efficiency for “always aware” applications

April 13, 2023, China - STMicroelectronics has launched three accelerometers with built-in advanced processing engines that enhance the sensor's ability to work autonomously, allowing the system to respond faster to external events while reducing power consumption.

LIS2DUX12 and LIS2DUXS12 use STMicroelectronics' third-generation MEMS manufacturing process and add new programmable features, including machine learning core (MLC), advanced finite state machine (FSM) and enhanced pedometer. Another entry-level accelerometer, the LIS2DU12, can be used for less demanding applications. All three products are equipped with the latest industry-standard I3C interface, general-purpose digital functions for detecting events, and anti-aliasing filters for high accuracy at low sampling frequencies, enabling operation with negligible power consumption. Accurate gesture detection.

LIS2DUX12 和 LIS2DUXS12 集成的MLC使人工智能 (AI) 算法能够可靠地检测人体活动，同时FSM进一步增强了活动识别的性能。两个功能相辅相成，让传感器能够自主处理数据，降低主处理器与传感器的交互操作和数据处理工作量，显著降低功耗，提高系统响应速度。此外，通过部署自适应自配置 (ASC) 功能，加速度计可以独立调整自己的设置，例如，量程和频率，以进一步优化测量性能，充分利用每一毫安的功耗。

LIS2DUXS12 also features STMicroelectronics' unique Qvar® sensing channel, which can sense changes in the surrounding electrostatic environment and provide presence detection and proximity detection functions. This feature enables developers to increase the value of biometric sensing applications such as user interface control, liquid detection, heart rate monitors and more. In user interface applications, Qvar® is combined with acceleration signals to eliminate potential false positives in double-click and multi-click events.

The new smart accelerometers provide context awareness for advanced wearables, true wireless stereo (TWS) speakers and headphones, smartphones, hearing aids, game controllers, smart watches, asset trackers, robots, and IoT devices. All three products use STMicroelectronics' latest ultra-low-power architecture, which combines inherent ultra-low-power technology with anti-aliasing filters to improve application performance and eliminate noise in signals. Users can obtain ready-made MLC and FSM algorithms through STMicroelectronics' MEMS GitHub model library to implement complex gesture detection, asset tracking and other use cases.

With these enhancements, LIS2DUX12 and LIS2DUXS12 can empower the next generation of "onlife" applications, intuitively and seamlessly supporting daily activities, always online, always aware, and continuously synchronized with user needs. The new products expand STMicroelectronics’ portfolio of AI-enhanced MEMS sensors, which includes inertial measurement units (IMUs) launched in 2019.

The LIS2DUX12 and LIS2DUXS12 are in production now and are available in a 2mm x 2mm x 0.74mm 12-pin LGA package.