Aug 4, 20234380
The concept of artificial intelligence was first proposed in the 1950s, more than 60 years ago. However, it is not until recent years that artificial intelligence has ushered in explosive growth. It is due to the increasingly mature Internet of Things, big data, cloud computing and other technologies.
The Internet of Things enables a large amount of data to be obtained in real time. Big data provides data resources and algorithm support for deep learning. Cloud computing provides flexible computing resources for artificial intelligence. The organic combination of these technologies drives the continuous development of artificial intelligence. This field has achieved substantial progress.
Product defect detection
Due to the application of deep learning, the process of defect detection in manufacturing production lines is getting smarter. The deep neural network integration allows the computer system to identify surface defects such as scratches, cracks, leaks, etc.
The detection is done by data scientists training a visual inspection system with a given defect detection task. A deep learning-driven inspection system, combined with a high optical resolution camera and GPU, creates a perception capability beyond traditional machine vision.
New detection techniques include synthetic data, transfer learning, and self-supervised learning, among others. Among synthetic data, Generative Adversarial Networks (GAN) data-generating tools examine images that quality inspectors consider “normal” and synthesize defective images for use in training AI models. At the same time, transfer learning and self-supervised learning are used to solve specific problems. As data accumulates, defect detection algorithms become more precise.
There are many sorting operations in the manufacturing industry. When talking about manual operations, the speed is slow, and the cost is high. Suitable working temperature environment needs to be provided. If industrial robots are used for intelligent sorting, the cost can be greatly reduced, and the speed can be increased.
Let's take sorting parts as an example. The parts that need to be sorted are often not neatly arranged. Although the robot has a camera to see the parts, it does not know how to pick them up successfully. In this case, machine learning technology is useful.
First, let the robot perform a random sorting action, and then tell it whether the action successfully picked up parts or caught empty. After many times of training, the robot will have a higher success rate. The success rate will be higher. After a few hours of learning, the robot's sorting success rate can reach 90%, which is comparable to that of skilled workers.
Warehouse Management and Logistics
The company replaces the two workstations that work three shifts a day with a robot. The robot is equipped with a machine vision system, and the RFID code can be scanned for sorting orders and delivery places. The judgment of finished products, empty boxes, and waste products is gradually improved by AI learning algorithms. Recognition rate, the initial recognition rate was only about 62%. Each shift needs to cooperate with a worker to fill in the gaps. With the accumulation of data, the AI recognition model has been continuously improved.
After 9 months, the comprehensive recognition rate has increased to 96%. Finished product recognition is completely accurate, and there is no need to keep people in the warehouse to fill in the gaps. It is only necessary to pick out a very small number of empty boxes when recycling waste.
Now there are more personalization, but the cost of personalized production is huge. The only way is mass customization, which uses personal consumption data to analyze and form a comprehensive order, and then distributes it on the platform for mass production to reduce the finished product In terms of unit price. However, although e-commerce has a large amount of consumer behavior data, the data always lags behind the actual demand. In this application scenario, the accuracy of the analysis platform needs to be maximized to increase.
Remote operation and maintenance service
The remote operation and maintenance platform conducts real-time monitoring of key parameters of the production process and production equipment through technologies such as the Internet of Things, big data and artificial intelligence algorithms, and timely alarms for faults.
Functions such as predictive maintenance and auxiliary decision-making supported by industrial big data analysis and artificial intelligence algorithms can further reduce business trips and shutdown delays caused by unplanned shutdowns, enabling industrial enterprises to achieve less-manned, unmanned, and efficient operation and maintenance.