AI Cause Estimation for ATM Banknote Handling Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The identification of errors in paper sheet handling devices, such as banknote jamming, often relies heavily on the skill of maintenance workers, leading to prolonged downtime and compromised service.
Innovation Solution
A cause estimation device and system that utilizes artificial intelligence to analyze sensor data from ATMs, generate learning models, and provide automated error diagnosis and recovery procedures, enabling remote maintenance support.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automated cause estimation system is implemented, then service reliability is improved, but device complexity increases
Solution Approach 1:
The patent introduces a cause estimation device as an intermediary system that collects sensor data from the paper sheet handling device, analyzes it using AI models, and provides cause estimates to maintenance workers. This mediator handles the complexity of AI analysis externally, allowing the core paper sheet handling device to remain relatively simple while still achieving improved reliability through automated diagnostics.
2Loss of time
If AI-based automated diagnosis is used, then loss of time is reduced, but device complexity increases
Solution Approach 1:
The system performs preliminary actions by collecting sensor data continuously and pre-processing it into feature quantities before actual errors occur. When an error happens, the cause estimation model can immediately analyze the pre-processed data and provide rapid diagnosis. This preliminary preparation reduces the time needed for actual error analysis without requiring complex real-time processing during critical failure moments.
3Measurement precision
If detailed sensor data collection is implemented, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent extracts only the necessary and relevant sensor data features that are actually useful for cause estimation, rather than collecting and processing all possible sensor data. The system identifies key feature quantities from sensor readings that correlate with specific error causes, extracting only these critical parameters for analysis. This selective extraction maintains high measurement precision for error detection while avoiding the complexity of processing exhaustive sensor datasets.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A cause estimation device (3) includes a collection unit (21A), a learning unit (21B), and an estimation unit (21C). The collection unit collects status data for each unit in an ATM (2) that handles banknotes. The learning unit generates a learning model for estimating a cause related to an error or an error symptom of the ATM for each unit with respect to the error or the error symptom, based on the status data. Upon detecting the error or the error symptom from the status data, the estimation unit estimates the cause for each unit, based on the learning model, and notifies a management device (4) of the cause, which is a result of the estimation. Upon receiving the cause from the cause estimation device, the management device notifies the worker terminal (6) and the maintenance company terminal (5) of the recovery procedure corresponding to the cause. The service deterioration of the ATM due to a prolonged recovery time can be avoided.