Six-axis IMU features AI core for intelligent motion signal processing at the edge
The STMicroelectronics LSM6DSV16X IMU combines an accelerometer and gyroscope with a machine learning core in a single compact package, to enable context awareness and configurable motion tracking without recourse to a host controller.
STMicroelectronics has introduced the LSM6DSV16X, a six-axis inertial measurement unit (IMU) which features advanced artificial intelligence (AI) capabilities to bring smart motion-sensing capabilities to IoT devices at the edge of the network.
The device is ideal for functions such as motion tracking and gesture detection, indoor navigation, image stabilization, and vibration monitoring and compensation.
The LSM6DSV16X includes standard IMU components: a three-axis digital accelerometer and a three-axis digital gyroscope. But on top of this, ST has built in a triple core for processing acceleration and angular rate data on three separate channels, the user interface, optical image stabilization, and electronic image stabilization, with dedicated configuration, processing, and filtering functions.
The LSM6DSV16X’s AI capabilities also enable it to apply edge computing functions to the raw motion data that it captures. Its finite state machine (FSM) performs configurable motion tracking, while a machine learning core (MLC) supplies context awareness, and provides exportable AI features for IoT applications.
The LSM6DSV16X supports the adaptive self-configuration feature, which allows the FSM to automatically reconfigure the device in real time based on the detection of a specific motion pattern. Reconfiguration can also be driven by the output of a specific decision tree configured in the MLC, without any intervention from the host processor.
The LSM6DSV16X embeds ST’s new Qvar function, which detects variation in electric charge, for user interface functions such as tap, double-tap, triple-tap, long press, or L/R or R/L swipe.
The LSM6DSV16X features an analog hub which can interface to an external analog input and convert it to a digital signal for processing.