AI on Microcontrollers

Application of e-AI in a semiconductor factory

A demonstration experiment on equipment abnormality detection and predictive maintenance by e-AI was conducted at the semiconductor front-end factory, Renesas Naka, in Hitachinaka City, Ibaraki Prefecture, Japan. New added value, such as abnormality detection and predictive maintenance, can be realized without major remodelling to existing facilities, while utilizing the existing facilities. In addition, e-AI has greatly improved detection precision with regard to facility abnormalities that previously could only be properly judged by a skilled technician or operator. Following this experiment, Renesas received inquiries from more than 40 customers in Japan and overseas and started discussions with more than 10 AI-related partners for concretization of business use-cases.

Matchmaker+ Anbieter zum Thema

zu Matchmaker+

This example convinced customers that e-AI could contribute to their business and could solve social problems. At the Naka factory, preparations are now under-way for practical use of e-AI for abnormality detection and predictive maintenance.

Hardware-accelerated AI

Renesas intends to continue advancing its range of e-AI solutions based on DRP technology (Dynamically Reconfigurable Processor). Renesas’ exclusive DRP allows real-time image processing at low power consumption. It is a hardware intellectual property (IP) module capable of dynamically changing the configuration of its processing circuit from one clock cycle to the next. DRPs have been in mass production use for more than a decade in applications such as broadcasting equipment and digital cameras.

This exclusive Renesas technology combines the performance of hardware with the flexibility of software, adding value to a wide range of products. Moving forward, Renesas plans to extend the use of DRP to embedded AI, delivering excellent power performance and unmatched flexibility tracking the ongoing advances in neural networks. After the launch of the RZ/A2M in September 2018, Renesas will release in the second half of 2019 an AI accelerator incorporating a DRP with AI processing performance 100 times faster than software processing, and the ability to perform real-time inference execution through embedded AI. In 2021, Renesas will release a next-generation AI accelerator 1,000 times as powerful as a conventional MPU. These Renesas AI innovations will add value to embedded devices and usher in the possibility of endpoints capable not only of inference execution but of learning as well.

The Author

Steve Norman

is Manager Global Ecosystems with Renesas Electronics and has extensive experience in introducing new technologies in various areas, particularly in the smart energy market where he advised the UK government. His current role is to improve the customer experience with the solutions from Renesas.

1. AI on Microcontrollers
2. Application of e-AI in a semiconductor factory