ST MEMS Sensor Roadmap Revealed, Edge Computing Enhanced
Real-time sensor data deep learning via ISPU embedded
[Editor's Note] If semiconductors are the rice of industry, MEMS are the beans and grains of industry, being installed in various products and increasing added value. Just like various types of grains, MEMS sensors are widely used in industry, taking on various roles such as acceleration, gyro, inertia, geomagnetism, pressure, and temperature. Recently, the trend is to develop MEMS sensor functions through edge computing reinforcement and minimizing power consumption through ISPU. It is expected that the integration of sensors and AI will lead to functional innovation and promote the era of Onlife, which implements the fusion of technology and the physical world. Accordingly, we met with Ji-Jun Young, Vice President of STMicroelectronics, to hear about the ST MEMS sensor roadmap and ISPU.
▲Ji Jun-young, ST Microelectronics Vice President
■ What is a Mems sensor and what are its main applications? MEMS is an abbreviation for Micro-Electro Mechanical Systems, which means an ultra-small precision mechanical system. It is a system that has a fine silicon-based three-dimensional structure and handles various input and output signals. ST's MEMS sensors are used in a wide range of fields, including portable electronic devices, home appliances, game consoles, IoT, medical, automotive, and industrial systems.
■ What are the major market demands for Mems sensors recently? With the increasing demand for battery-powered mobile devices, ST is introducing a line of MEMS sensors that can kill two birds with one stone: high performance and low power consumption.
■ What solutions does the ST product line have for this? ST is a global leading manufacturer of MEMS sensors. △We provide almost all types of MEMS sensor products, including acceleration sensors for measuring acceleration, gyro sensors for measuring rotational speed, inertial sensors with both acceleration and gyro sensors embedded, geomagnetic sensors for measuring magnetism, pressure sensors for measuring atmospheric pressure, and MEMS microphones for measuring voice.
Recently, we announced a roadmap to strengthen edge computing through ISPU. We are preparing solutions that can minimize power consumption at the sensor and system level.
■ Please explain what ISPU (Intelligent Sensor Processing Unit) is. The ISPU is a C-programmable DSP embedded inside the ST MEMS sensor, supporting 32kbytes of program RAM, 8kbytes of data RAM, and FPU operations.
■ The benefits provided by ST’s ISPU are: Signal processing on real-time sensor data and execution of software algorithms such as machine learning and deep learning can be done inside the sensor, which can speed up decision-making by reducing data transmission outside the sensor. And by keeping data local, privacy can be protected, while minimizing the sensor data processing role of the main MCU, reducing the power consumption of the entire system.
■ Lastly, a word to the readers We hope that this webinar will be an opportunity to introduce ST MEMS sensors and ISPU functions to many people. We ask for your interest and participation.
thank you