ICEST 2025 Lecture - TinyML and IIoT Based Product Quality Classification for Food Industry

#artificial-intelligence #automation #edge-computing #control

An important problem that most industrial systems face is requirement for real-time processing of large amounts of data and a rapid inference response with minimal latency. Simultaneously, in widely adopted artificial intelligence/machine learning (AI/ML) implementation in industrial control and robotics, there is a clear recent shift from cloud computing to edge computing due to safety, reliability and other aspects. This is challenging, as small edge devices have limited hardware capabilities to process large amounts of data within limited timeframes and for deployment and execution of large machine learning models. The paper introduces an approach based on TinyML (Tiny Machine Learning), an emerging branch of machine learning that allows models to run on small, resource constrained devices, for a product classification problem in the food industry. The proposed solution also relies on IIoT (Industrial Internet of Things) for connecting industrial devices. Compared to our previously published results for similar problems which are based on powerful embedded platforms, the approach adopted here demonstrates an attempt to utilize affordable, lightweight hardware as an alternative, while still maintaining required performance. The proposed system is suitable for edge computing tasks in food industry applications due to its low power consumption, availability and cost effectiveness. The suggested approach allows for prompt anomaly detection and advances automation efforts in practical industrial settings.

• Add Event to Calendar
  iCal
  Google Calendar

• Ohrid, Macedonia
• Macedonia

---

• Contact Event Host
• Co-sponsored by IEEE Republic of North Macedonia Section

• Starts 17 June 2025 10:00 PM UTC
• Ends 26 June 2025 10:00 PM UTC
• No Admission Charge

•