Intended for edge-AI applications, the new sensor device enables extra features and longer battery runtime in consumer wearables, asset trackers, and impact and fall alarms for workers.
The LSM6DSV32X extends the family of smart sensors that contain ST’s machine-learning core (MLC) with AI algorithms based on decision trees. With the MLC for context sensing and a finite state machine (FSM) for motion tracking, these sensors will help product developers add new features, minimise latency, and save power.
Leveraging the embedded features the LSM6DSV32X has been able to significantly reduce the power budget for functions such as gym-activity recognition to below 6µA. The sensor also embeds ST’s Sensor Fusion Low-Power (SFLP) algorithm to perform 3D orientation tracking at just 30µA. And by supporting adaptive self-configuration (ASC), the module autonomously reconfigures sensor settings in real-time to continuously optimize performance and power.
In addition to the accelerometer and gyroscope, the LSM6DSV32X integrates ST’s Qvar electrostatic charge-variation sensing to handle advanced user-interface functions such as touching, swiping, and tapping. The module also contains an analogue hub for acquisition and processing of external analogue signals.
Product developers can rely on a large selection of ready-to-use libraries and tools to accelerate time to market for new products. These include the intuitive MEMS Studio environment, which supports evaluation and use-case development, and a dedicated GitHub repository that provides code examples such as sports-activity and head-gesture recognition.
Resources also include hardware adapters for connecting the IMU to ST’s evaluation and proof-of-concept boards such as the ProfiMEMS board, Nucleo sensor expansion board, and Sensortile.box PRO.
The LSM6DSV32X is scheduled to enter volume production in May 2024 in a 2.5mm x 3mm x 0.83mm 14-lead LGA package.
