AI at the edge enables real-time failure prediction with no cloud server required
The new AI Learning Edge chip on ROHM’s device
ROHM’s AI chip concept
Santa Clara, Calif. and Kyoto, Japan, Nov. 29, 2022 (GLOBE NEWSWIRE) — ROHM Semiconductor today announced that it has developed an on-device learning solution AI chip (SoC with on-device learning AI accelerator) for edge computing terminals in IoT field. The new AI chip uses artificial intelligence to predict failures (predictive failure detection) in electronic devices equipped with real-time motors and sensors with ultra-low power consumption.
Typically, AI chips perform learning and inference to perform artificial intelligence functions because learning requires a large amount of data to be captured, compiled into a database, and updated if necessary. Thus, the AI chip that performs the learning requires significant computing power which necessarily consumes a large amount of energy. Until now, it has been difficult to develop AI chips that can learn in the field with low power consumption for peripheral computers and terminals to create an efficient IoT ecosystem.
Based on an “on-device learning algorithm” developed by Professor Matsutani of Keio University, ROHM’s newly developed AI chip mainly consists of an AI accelerator (AI dedicated hardware circuit) and ROHM’s “tinyMicon MatisseCORE™” high-performance 8-bit processor. The combination of the ultra-compact 20,000-gate AI accelerator with a high-performance processor enables learning and inference with ultra-low power consumption of only tens of mW (1000 times smaller than conventional AI chips capable of learning). This enables real-time failure prediction in a wide range of applications, as “anomaly detection results” (anomaly score) can be generated numerically for unknown input data at the site where the equipment is installed without involving a cloud server.
In the future, ROHM plans to integrate the AI accelerator used in this AI chip into various IC products for motors and sensors. Marketing is expected to begin in 2023, with mass production expected in 2024.
Professor Hiroki Matsutani, Department of Information and Computing, Keio University, Japan
“As IoT technologies such as 5G communication and digital twins advance, cloud computing will need to evolve, but processing all data on cloud servers is not always the best solution in terms of load, cost and energy consumption. With the “on-device learning” we pursue and the “on-device learning algorithms” we have developed, we aim to achieve more efficient data processing on the edge side to build a better IoT ecosystem. Through this collaboration, ROHM has shown us the way to commercialization in a cost-effective way by advancing on-device learning circuit technology. I expect the prototype AI chip to be incorporated into ROHM’s IC products in the near future.
About tinyMicon MatisseCORE
tinyMicon MatisseCORE (Matisse: Mmicrophone acalculation unit for ofthe ssize withsequencer) is ROHM’s proprietary 8-bit processor developed with the goal of making analog ICs smarter for the IoT ecosystem. An instruction set optimized for embedded applications, combined with the latest compiler technology, provides fast arithmetic processing in a smaller chip area and program code size. High-reliability applications are also supported, such as those requiring qualification to ISO 26262 and ASIL-D vehicle functional safety standards, while the proprietary on-board “real-time debugging function” prevents the debugging process to interfere with the operation of the program, allowing debugging to be performed while the application is running.
Detail of ROHM’s AI chip (SoC with on-device learning AI accelerator)
The prototype AI chip (Prototype Part No. BD15035) is based on an on-device learning algorithm (three-layer neural network AI circuit) developed by Professor Matsutani of Keio University. ROHM shrunk the AI circuit from 5 million gates to just 20,000 (0.4% of the size) to reconfigure it for commercialization as a proprietary (AxlCORE-ODL) AI accelerator controlled by ROHM’s high-efficiency 8-bit processor, tinyMicon MatisseCORE, which enables AI learning and inference with ultra-low power consumption of just tens of mW. This makes it possible to digitally output “anomaly detection results” for unknown input data patterns (i.e. failure prediction (detection of predictive signs of failure) by onsite AI while maintaining maintenance costs. low communication and cloud server.
To evaluate the AI chip, ROHM offers an evaluation board equipped with Arduino compatible terminals which can be equipped with an extension sensor board for connection to an MCU (Arduino). Wireless communication modules (Wi-Fi and Bluetooth®), as well as a 64kbit EEPROM memory, are mounted on the board. By connecting units such as sensors and attaching them to the target equipment, it will be possible to check the effects of the AI chip from a screen. This evaluation board will be loaned by ROHM Sales. Please contact ROHM sales for more information.
AI Chip Demo Video
A demo video showing this AI chip used with the evaluation board is available here: https://youtu.be/SVn5CKFX9Uo
tinyMicon MatisseCORE™ is a trademark or registered trademark of ROHM Co., Ltd.
Bluetooth® is a trademark or registered trademark of Bluetooth SIG, Inc.
 On-device learning: running training and inference on the same AI chip
CONTACT: Travis Moench ROHM Semiconductor 858.625.3600 email@example.com Heather Savage BWW Communications 720.295.0260 firstname.lastname@example.org
#ROHM #developing #ultralowpower #artificial #intelligence #chip #device