Smart AWS Industrial Machine Learning Technology
Inspection in the manufacturing industry requires a series of controls such as testing, measuring, and examination of an object, material or activity. These controls are done by technologies or people, or a combination of both with the purpose of determining whether or not it is in proper condition. In this blog, I’ll explain the 4 industrial machine learning services provided by AWS that helps manufacturers to keep their operations running within specified limits and ensure quality and prioritize safety.
#1 — AWS Monitron
AWS Monitron is a combination of sensors, gateway, and machine learning that allows industries to monitor end-to-end equipment in their production line. Manufacturers can connect these sensors produced by AWS to their legacy equipment to measure the temperature and vibration, etc.
How does it work?
1- These smart sensors are Bluetooth low energy type, are attached to the equipment to measure the temperature and vibrations, and then send these data to the Gateway.
2- Gateway is a physical device connected via wifi to the local Internet and automatically gets data from all sensors connected to it. Then this data is sent to the AWS cloud.
3- All data sent by the gateway are stored in AWS, using a machine learning model to identify all anomalies and defects detected by the sensors and automatically sent the result to a mobile application.
4- A friendly dashboard is displayed with a list of all the equipment in the mobile application, which a technical person can identify whether the equipment is healthy or unhealthy.
5- The Quality Control Officer identifies all the sensors that are unhealthy and then can go and do some inspections where unhealthy equipment is set up.
#2 — AWS Panorama Appliance
AWS Panorama Appliance is a smart device that allows customers to make predictions based on the images captured by cameras that are added to this device. This device has the capability to use computer vision to improve quality control.
How does it work?
1- Cameras type H.264 for live streaming sending the data to Appliance.
2- This device supports up to 10 cameras simultaneously, collects all data, and then sends the data to the cloud.
3- Machine Learning makes a prediction based on the data collected by Appliance.
#3 — Amazon Lookout for Vision
Amazon Lookout for Vision is a Machine vision intelligence for product defect inspection based on deep learning. This enables customers to monitor and find defects in their product line in real-time.
The following image shows how Machine Learning makes a prediction based on these images and visions.
#4 — Amazon Lookout for Equipment
Amazon Lookout for Equipment allows customers to use their own sensors to detect anomalous equipment, it’s quite the same logic as Amazon Monitron, but in this case, customers are using their own sensors instead of AWS’s. In addition, this service is based on machine learning that trains the data collected by these sensors and performs predictive maintenance.
1- Sensors identify the anomalies in legacy equipment and send the data to the cloud.
2- Analyze data from inputs like generators, compressors, turbines and, so on.
Conclusion
All these smart industrial equipment are very useful in the industrial world today, helping customers to automate all processes and reduce costs. There are simple cost-effective solutions leveraging machine learning capabilities to make predictions in real time.
