New Inertial Sensor Module from STMicroelectronics Enables In-Sensor AI Training | 黑森爾電子
聯繫我們
SalesDept@heisener.com +86-755-83210559 ext. 809
Language Translation

* Please refer to the English Version as our Official Version.

New Inertial Sensor Module from STMicroelectronics Enables In-Sensor AI Training

Technology Cover
發佈日期: 2022-06-28, STMicroelectronics

STMicroelectronics (ST;), the world's leading semiconductor company serving multiple electronic applications, has launched a new inertial sensor with a built-in Intelligent Sensor Processing Unit (ISPU), promoting the arrival of the onlife era, where people and the Trained smart devices interact, and smart technologies move from the network edge to deep edge devices.

The ISM330ISN is an always-on 6-axis inertial measurement unit (IMU) sensor with smart technology embedded inside that delivers best-in-class measurement performance and accuracy in terms of size and power consumption. ST's new IMUs are ideal for IoT and industrial applications, speeding up response times and extending device battery life for industrial applications such as predictive maintenance condition monitors, battery-powered asset trackers, robotics, and more.

The intelligence built into the ISM330ISN enables smart devices to perform advanced motion detection functions in the sensor without interacting with an external microcontroller (MCU), reducing system-level power consumption. ST's approach is to integrate a dedicated processor ISPU optimized for machine learning applications directly on the sensor chip. This smart core occupies a very small chip area, so the package area of ​​the ISM330ISN module is smaller than that of a typical sensor that integrates an MCU within the package. 50% smaller solution and 50% lower power consumption.

Developers can program the ISPU with NanoEdge AI Studio, ST's market-proven AI tool, which many customers have used to deploy AI applications on STM32 microcontrollers. Now, the technology can be used to program ISPUs, allowing users to easily generate automatically optimized machine learning libraries. Requires very little data, and with just a few mouse clicks, an anomaly detection library with AI learning capabilities can be designed directly within the ISPU. Developers do not need to have specialized knowledge and skills in data science.

"Intelligence, once deployed on application processors at the edge of the network, is now moving inside sensors at the edge of depth. The ISM330ISN IMU heralds a new class of smart sensors," said Simone Ferri, general manager of marketing for the Analog MEMS and Sensors Group at STMicroelectronics. Coming, starting to leverage embedded AI for complex tasks, such as pattern recognition and anomaly detection, greatly improves system energy efficiency and performance.”

While demonstrating the new creative opportunities brought about by AI, the ISM330ISN also features the same 3mm x 2.5mm x 0.83mm package size as traditional inertial modules. As a result, designers can quickly and cost-effectively upgrade products without changing existing board layouts.

The ISM330ISN is part of ST's 10-year Product Lifetime Guarantee program, which provides product designers and manufacturers with a long-term supply guarantee.

Detailed technical information

The ISM330ISN, a member of ST's iNEMO IMU inertial measurement unit family, integrates a three-axis accelerometer and a three-axis gyroscope with low noise sensing performance and an output data rate (ODR) of 6.6kHz. With the ISPU smart core, the sensor's high-accuracy performance is stable while consuming only 0.59mA in combined mode with both accelerometer and gyroscope active.

Based on a digital signal processing (DSP) architecture, ST's ISPU is extremely compact and energy-efficient, has 40KB of RAM memory, and occupies only 8000 logic gates on the sensor chip. The ISPU performs single-precision floating-point operations and is ideal for designing machine learning applications and binary neural networks.


相關產品