Taiwan manufacturer adopts end-point AI sensors

  • August 7, 2022
  • William Payne

Taiwanese IoT AI specialist aiSensing has developed an endpoint AI-based vibration sensor for a large specialty manufacturer in Taiwan. Believed to be Nan Pao, the customer is one of Taiwan’s largest manufacturers and is a major global producer of adhesives and liquid and powder coatings, with factories in China, Vietnam, Indonesia, Thailand, Malaysia and India. The vibration sensor monitors equipment status and is able to identify multiple different fault modes.

The solution is based on an EOS S3 multi-core sensor processing SoC and SensiML Analytics from San Jose based endpoint AI specialist SensiML, a subsidiary of QuickLogic.

aiSensing’s system incorporates AI/ML technology to monitor the status of manufacturing equipment locally without the need for an Internet-based cloud connection.

Since the AI implementation is local, rather than cloud-based, the system features low cost, low latency and fast reaction times while simultaneously providing higher data security.

A leading manufacturer of specialty adhesives, footwear adhesives, hot-melt adhesives, and liquid and powder coatings, the company has adopted an edge AI-based approach to detect anomalies for vacuum pumps and chilling machines used in its manufacturing flow, including problems related to lack of lubrication, water leakage, bearing failures, and belt failures.

The AI-based endpoint was developed on a QuickLogic EOS S3 ultra-low power multi-core Arm Cortex MCU-based SoC, which delivered more than enough processing bandwidth for the application at a low cost. The AI application running on the QuickLogic device was built using the SensiML Analytics Toolkit, which provided a complete solution for the quick development of this sophisticated IoT endpoint.

“Smart manufacturing is a significant trend across a broad range of industries,” said Chris Rogers, chief executive officer at SensiML. “Predictive maintenance is one of the core initiatives in that trend, and aiSensing’s vibration sensor is a great example of how to effectively use AI to implement a practical and cost-effective predictive maintenance solution.”

“Our endpoint AI-based vibration sensor has been very successful,” said Dennis Chu, chief technology officer at aiSensing. “Its low power consumption, fast response times, and low cost are the ideal combination of features for this predictive maintenance application. With the SensiML tools, we can easily modify the design to address new and unique requirements for our customers.”