Sensors deliver machine learning core efficiencies
STMicroelectronics is extending the benefits of its motion-detection machine-learning core (MLC) technologies into industrial and high-end consumer applications with the latest ISM330DHCX and LSM6DSRX six-axis iNEMO inertial measurement units (IMU).
The MLC performs basic AI pre-processing of motion data using about 0.001 times the power a typical microcontroller (MCU) would consume to complete the same task. As a result, IMUs featuring this IP can offload the host MCU, enabling longer battery runtime, lower maintenance, and reduced size and weight in context-aware and motion-sensing devices.
Following the launch of the first MLC-enhanced commercial IMU last year, ST’s LSM6DSRX and ISM330DHCX are ready for high-end consumer and industrial applications such as augmented/virtual reality, drone flight controls, dead-reckoning navigation systems, disk-antenna positioning systems, fleet management, container-tracking devices, and dynamic inclinometers for industrial vehicles.
The consumer-grade LSM6DSRX contains a three-axis accelerometer and a three-axis digital gyroscope with extended full-scale angular-rate range up to ±4000dps and leading-edge performance in temperature and time. The industrial-grade ISM330DHCX comes with ten-year product-longevity assurance and is specified from -40 to +105°C, with embedded temperature compensation for superior stability.
In each device, the MLC interacts with integrated finite-state machine (FSM) logic, which can run simple repetitive algorithms like counting steps, hits, or rotations at lower power than is possible in a microcontroller. The FSM signals to the main controller after detecting a preset number of events or after a defined time has elapsed.
Both devices are in production now. The ISM330DHCX is available as a 14-lead plastic land grid array (LGA) device, priced from $4.00 in quantities of 1000 pieces. The consumer-grade LSM6DSRX is $3.50.