Quality is one of the main criteria on a production line and the ability to ensure product quality during the full production process is essential. In this use case, we will look at an example using a roasting machine. Roasting is a delicate craft that requires precision, expertise and consistency to achieve the desired flavor profiles in coffee, nuts and other food products. Traditional roasting machines have been the backbone of the industry, but now, with the integration of technologies such as Artificial Intelligence (AI), we are witnessing a remarkable leap forward in the production quality. By monitoring variables such as temperature, humidity or roasting duration, AI can continuously analyze and adjust the roasting process in real-time to optimize the quality of the final product with unparalleled precision and consistency.

Approach

This use case is based on the "Production quality" dataset from Kaggle.
The goal was to predict roasting quality based on different variables.
The roasting machine consists of 5 chambers containing 3 temperature sensors each. There are also sensors used to measure the layer height and the humidity of the raw material entering the machine. So, 17 sensors in total.
In this example, the quality is measured in a laboratory. The AI model will use examples to understand the relationship between the sensor values and the quality measured. This replaces the manual steps that take place in the laboratory to gain time.
The dataset contains sensor data recorded every minute (from all 17 sensors), and a quality value determined each hour. For simplicity, we only look at the sensor data that precedes the quality measurement. Another approach would be to concatenate all sensor measurements.
We then used NanoEdge AI Studio to create an Extrapolation project that can predict the quality of the roasted goods based on the data from the 17 sensors every hour.

Sensor

Temperature, humidity and displacement sensors.

Data

Extrapolation target Production quality
Signal length 17 (multi sensors)
Data rate every hour

Results

Extrapolation:
90.81% accuracy, 0.1 Kbytes of RAM, 189.8 Kbytes of Flash memory

uc-neais-production-quality-result uc-neais-production-quality-result uc-neais-production-quality-result
Model created with
NanoEdge AI Studio
NanoEdge AI Studio
Compatible with
Any STM32 MCU
Any STM32 MCU

Resources

Model created with NanoEdge AI Studio

A free AutoML software for adding AI to embedded projects, guiding users step by step to easily find the optimal AI model for their requirements.

NanoEdge AI Studio NanoEdge AI Studio NanoEdge AI Studio

Compatible with Any STM32 MCU

The STM32 family of 32-bit microcontrollers based on the Arm Cortex®-M processor is designed to offer new degrees of freedom to MCU users. It offers products combining very high performance, real-time capabilities, digital signal processing, low-power / low-voltage operation, and connectivity, while maintaining full integration and ease of development.

Any STM32 MCU Any STM32 MCU Any STM32 MCU

You might also be interested by

Predictive maintenance | Accelerometer | NanoEdge AI Studio | Video | Partner | Industrial

Anomaly detection with on-device learning with Rtone

Anomaly detection solution on industrial equipment, running on STM32 MCU.

Object detection | Vision | STM32Cube.AI | Idea | GitHub | Video | Dataset | Smart building

Personal protective equipment detection

Detection of personal protective equipment on workers using an object detection AI model.

Smart city | Smart home | Smart office | Context awareness | Vision | AI for Linux | Idea | Video

People detection and counting solution

Optimized computer vision using an MPU running at 8 FPS.