Edge Instrument Anomaly Detection for Repetitive Task Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional defect detection techniques in manufacturing environments are inadequate for identifying anomalies in repetitive tasks, leading to potential quality issues in assembled items.
Innovation Solution
The implementation of machine learning-based anomaly detection using sensor data from edge instruments, such as tools and gloves with embedded sensors, to identify deviations from expected sequences of actions in repetitive tasks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional defect detection techniques are used at the end of assembly process, then manufacturing cost is reduced, but quality issues are not detected early in the assembly process
Solution Approach 1:
The patent applies preliminary action by implementing anomaly detection during the assembly process itself rather than waiting until the end. Sensors embedded in edge instruments collect data throughout the assembly process, and the machine learning model continuously monitors for deviations from expected patterns, enabling early detection and immediate correction of quality issues.
Solution Approach 2:
The patent uses an intermediary approach by introducing a machine learning model that acts as a mediator between the assembly process and quality detection. The model analyzes sensor data from edge instruments and identifies anomalies that conventional techniques would miss, bridging the gap between manufacturing operations and quality assurance.
2Measurement precision
If sensor data from edge instruments is collected and analyzed using machine learning models, then anomaly detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies self-service by embedding the machine learning model directly within the edge instrument itself. The instrument autonomously collects sensor data, processes it through the embedded model, and generates anomaly detections without requiring external computational resources. This self-contained approach improves detection accuracy while managing system complexity through integration.
Solution Approach 2:
The patent merges multiple functions into the edge instrument by combining sensors, machine learning model, and anomaly detection capabilities into a single integrated device. This consolidation improves measurement precision while reducing overall system complexity compared to having separate components for data collection, processing, and analysis.
3Reliability
If quality inspections are performed only at the end of assembly process, then manufacturing process speed is maintained, but defect detection capability is insufficient
Solution Approach 1:
The patent implements continuity of useful action by performing anomaly detection continuously throughout the assembly process rather than performing discrete inspections at the end. Sensors continuously collect data and the machine learning model continuously analyzes it, ensuring that quality monitoring occurs without interrupting the assembly workflow, thus maintaining productivity while improving defect detection capability.
Data Source
AI summary
Techniques are provided for machine learning-based anomaly detection for repetitive tasks performed using edge instruments. One method includes obtaining sensor data characterizing an orientation and/or an acceleration of an edge instrument utilized to perform a repetitive task, comprising a sequence of actions, by a user. The sensor data is obtained from at least one sensor embedded in the edge instrument. The sensor data is applied to a machine learning model trained to identify a deviation from an expected sequence of actions associated with the repetitive task. The machine learning model is embedded in the edge instrument. An automated action is initiated in response to the machine learning model identifying the deviation from the expected sequence of actions.