Source: The content is compiled from allabourcircuit
by Semiconductor Industry Observer (ID: icbank)
, thank you.
To drive innovation in microelectromechanical systems (MEMS), STMicroelectronics has launched an energy-efficient system that combines signal processing, artificial intelligence and MEMS sensors.
MEMS technology combines miniaturized electromechanical components such as microsensors and microactuators by integrating mechanical and electrical functions.
These devices are typically manufactured using a variety of industrial processes, including very large scale integration (VLSI) technology, micromachining, IC process sequences, etc.
An example of a MEMS sensor
Recently, in an effort to continue to push the boundaries of MEMS technology, STMicroelectronics has developed a
MEMS sensor that
integrates digital signal processing (DSP) and AI algorithms on the same silicon die.
In this article, we’ll explore how AI plays a role in MEMS devices, the challenges, and ST’s latest releases.
How does artificial intelligence scale on MEMS?
Researchers and designers are working around the clock
to integrate artificial intelligence into MEMS
to improve performance and expand the use cases of MEMS devices.
With this in mind, new
AI-MEMS architectures
are emerging.
One such architecture exploits resonators with nonlinear dynamics to enable machine learning (ML) processing in the mechanical domain.
Example of a MEMS device from the University of Sherbrooke
In 2018, researchers at the University of Sherbrooke in Canada
achieved a milestone
by
introducing reservoir computing
, which allowed MEMS oscillators to perform time series prediction and spoken word classification.
The researchers exploited the nonlinear dynamics of a silicon beam, which oscillates in space with a width 20 times thinner than a human hair.
The results of this oscillation are said to be used to construct a virtual neural network that projects the input signal into the higher-dimensional space required for neural network calculations.
This AI-MEMS simplifies the mechanical functions of the robot using accelerometers, which can generate control signals for the robot.
In general, fabricating scalable AI-MEMS architectures can accelerate the versatility of MEMS devices and signal processing.
In addition to providing improved performance, it can also eliminate the use of external microprocessors and field-programmable gate arrays (FPGAs).
When faced with challenges, MEMS designers
face several pitfalls and design limitations when manufacturing MEMS devices
.
Getting real data in smart MEMS sensors requires using higher resolution ADCs. Therefore, a 10-bit resolution ADC may not be helpful for certain applications such as health monitoring.
Additionally, when processing data for data transfer, limited bandwidth can create challenges and truncate data processing.
Designers also face challenges in manufacturing state-of-the-art MEMS sensors that integrate ML algorithms.
MEMS sensors that use classification algorithms such as support vector machines (SVM) require large memories to store large real-life data sets.
Despite the challenges of incorporating AI and ML into MEMS devices, ST hopes to make it easier.
Smart sensor processing unit meets MEMS sensor
To overcome all the potential challenges associated with building AI on MEMS, ST introduced
the Intelligent Sensor Processing Unit (ISPU)
, which integrates DSP with MEMS sensors on an IC.
The programmable DSP features a single-cycle 16-bit multiplier that can be easily operated with 16-bit variable-length instructions. It also includes a full-precision floating-point unit.
Overview of ST's ISPU
ISPU facilitates full-to-unit accuracy neural networks in quantized AI sensors.
With AI algorithms running on the DSP, the ISPU analyzes inertial data to improve the accuracy and efficiency of activity recognition and anomaly detection tasks.
Additionally, the ISPU supports edge AI computing, allowing the development of MEMS sensor algorithms using AI business models while maximizing ultra-low power consumption.
ST also claims that the product has the potential to reduce power by up to 80% while also reducing the size of system-in-package devices.
Speaking about the product, Andrea Onetti, executive vice president of STMicroelectronics' MEMS sub-group, said the new era - the "
Onlife Era
" - aims to advance sensor capabilities by reducing data transmission to speed up decision-making and enhance privacy by keeping data localized, while reducing size and power consumption, thereby reducing costs.
In summary, while achieving maximum privacy, Onlife Era aims to introduce MEMS devices that can sense real-life data, process complex AI algorithms, and take intelligent real-time actions.
★ Click
[Read original text]
at the end of the article
to view the original link of this article!
*Disclaimer: This article is originally written by the author. The content of the article is the author's personal opinion. Semiconductor Industry Observer reprints it only to convey a different point of view. It does not mean that Semiconductor Industry Observer agrees or supports this point of view. If you have any objections, please contact Semiconductor Industry Observer.
Today is the 2959th content shared by "Semiconductor Industry Observer" for you, welcome to follow.
Semiconductor Industry Observation
"
The first vertical media in semiconductor industry
"
Real-time professional original depth
Scan the QR code
, reply to the keywords below, and read more
Wafers｜ICs｜Equipment
｜Automotive Chips｜Storage｜TSMC｜AI｜Packaging
Reply
Submit your article
and read "How to become a member of "Semiconductor Industry Observer""
Reply
Search
and you can easily find other articles that interest you!
Click to read the original text to view this article
Original link!
京公网安备 11010802033920号