AI & Edge Computing
A system for scientifically and systematically analyzing and evaluating the results of detecting product packaging discrepancies based on big data and artificial intelligence (AI)
Field of AI application
Automate visual inspection using AI
Traditional inspection process -> Inspection personnel or automatic inspection equipment
It is difficult for inspectors to perform repeated inspection work, and it is difficult to manage uniform inspection standards among inspectors Automatic inspection devices are complex and time consuming to set up or adjust, and if the settings are strict Yield decreases, and if loose, defects increase and re-examination by the inspector is required
Defects such as damage, dents, stains, shape abnormalities, assembly abnormalities, etc. judged by the eyes of the inspector are used as filming equipment Image data is collected and judged, and the Gray Zone's defect judgment, which has been executed based on the inspector's intuition and experience, is made more precise by using artificial intelligence's Deep Learn
Autonomous learning of robot movements using AI
Unlike industrial robots that perform tasks by setting operating trajectories in advance In collaborative robots that aim to replace workers' work, in addition to pre-defined actions You need to learn flexible movements that are close to humans Obtain a baseline of behavior from the interaction of AI that implements appropriate movements and AI that judges them Use imitation learning to ensure safety and avoid dangerous behavior towards people
Analysis of causes of defects using AI
Reduce the loss rate by quickly stopping the process that causes defects and restarting them after taking countermeasures
AI extraction from sensor data when and where the problem process and operation that caused the occurrence of the defect occurred AI can detect anomalies in a short time
In the event of a defect, search past data and data are easily used to utilize related information, while related knowledge is utilized by search
When a defect occurs, the task of automatically creating and storing basic data to analyze the cause of the defect by leaving a record is automatically performed using AI to support the task to be performed by humans
AI-based facility maintenance
Application of techniques to learn the predictive patterns of failure and to predict and warn prior to failure
Install sensors for monitoring and predicting equipment operation status When the equipment starts operating, it can specify the operating state and abnormal state through sensing of temperature, pressure, vibration, sound, etc., and can be alerted according to the prediction pattern of operation by learning the AI
Proactively address down time due to equipment failures and prevent unexpected failures
Planned facility inspection to minimize and simplify personnel
Edge Computing
Key Features
Compute where data is actually collected
Edge AI Computing implements artificial intelligence in Edge Computing environments; centralized cloud computing facilities; or, rather than Offsite Data Center, compute near where data is actually collected
Processing data with or without an Internet connection
Smart decisions are made quickly without being connected to the Cloud or Data Center, and AI algorithms process data generated by devices regardless of whether they are connected to the Internet
True AI
Work based on machine learning (ML) models integrated within Edge devices can be implemented with strong artificial intelligence (True AI)
Reduce costs Reduce latency
Lower power usage, lower bandwidth, lower data privacy, higher security/scalability, lower latency, and more, making it a key technology in the future
Edge AI Computing Functional Features
Connectivity with existing facilities
- Most old and analog installations connect to the cloud or pre-process data Absence of the latest connectors available
- Edge Computer converts old equipment or its controls to MQTT, OPC UA or support the latest protocols such as Jetson
- Data can be formatted, manipulated, interpreted, or filtered directly from the Edge device, and only selected data Exchange with Cloud
- Simplify, efficient and fast cloud storage and reduce the bandwidth you need
Connectivity with multiple facilities
- Innovase Edge Computer collects data from multiple facilities through specialized control methods
- Manage local pre-evaluation and aggregation of machine data to quickly handle critical thresholds or production data
- Data is stored on site and visualized directly from Edge Computer
- Transfer to a higher-level cloud or database system
- Additional physical interfaces such as USB, DVI/DisplayPort, RS232 and more are also available
