STMicroelectronics' Latest Six-Axis IMU, the LSM6DSV32X, Packs a Machine Learning Core for TinyML

STMicroelectronics is expanding its range of edge artificial intelligence (AI) sensors, with integrated machine-learning core (MLC) accelerators, with the impending launch of the LSM6DSV32X six-axis inertial measurement unit (IMU).

"The LSM6DSV32X six-axis inertial module from STMicroelectronics has a large accelerometer full-scale range of 32g and 4,000 degrees-per-second (dps) gyroscope to measure intensive movements and impacts, including freefall height estimation," the company says of its latest smart sensor design. "Ready to drive future generations of 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 is one of a growing number of parts, which integrates STMicro's in-house machine learning core (MLC) accelerator — allowing it to run small on-device machine learning models, known as "tinyML," without the need to bother a host processor. Combined with a finite state machine for motion tracking, STMicro says this lets the new sensor deliver gym-related activity recognition at a power draw below 6µA — and to handle sensor fusion for 3D orientation tracking at just 30µA.

Like other parts in STMicro's smart-sensor range, the IMU includes the company's adaptive self-configuration system — allowing it to automatically reconfigure sensor settings on-the-fly to strike a balance between performance and power. The sensor also includes STMicro's Qvar electrostatic charge variation sensing system, to handle touch inputs including swiping and tapping without the need for additional hardware, and an analog hub that can acquire and process analog signals from external sensors.

STMicro has confirmed plans to start volume production of the LSM6DSV32X in May this year, with pricing starting at $2.98 per unit in 1,000-unit tray quantities. More information is available on the STMicro website.