Operator Profiling via Machine Vibration Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for operator profiling in manufacturing and factory control processes require additional sensors, which can inhibit natural movement and are costly or raise privacy concerns, and are limited in their ability to evaluate skill levels beyond pre-defined metrics or specific use cases.
Innovation Solution
A system that extracts segmented time series data from pre-installed sensor measurements using machine learning algorithms to evaluate operator skill levels without additional sensors, providing evaluation scores, identifying key movements, and recommending skill improvement actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If body-attached sensors are used to record operator motion, then motion data can be collected for evaluation, but the sensors inhibit natural movement of the operator
Solution Approach 1:
The patent extracts the measurement function from body-attached sensors and relocates it to the machine's existing sensors. The machine sensors capture operator-induced vibrations and movements indirectly, eliminating the need for physical sensors on the operator's body while preserving measurement capability
Solution Approach 2:
The patent introduces machine vibrations and sensor data as an intermediary between the operator's natural movement and the evaluation system. Instead of directly measuring operator motion, the system measures the machine's response to operator actions, providing indirect but effective measurement without physical contact
2Ease of operation
If video cameras are used to record operator motion, then natural movement is not inhibited, but additional sensors are required which increase cost and raise privacy concerns
Solution Approach 1:
The patent makes the machine's existing sensors serve multiple functions: their primary function for machine monitoring and their secondary function for operator skill evaluation. This eliminates the need for additional dedicated sensors while enabling operator profiling capabilities
Solution Approach 2:
The machine's existing sensor system serves itself by simultaneously performing both machine health monitoring and operator skill evaluation. The same sensors that detect machine status also capture operator-induced vibrations, making the system self-sufficient for dual purposes
3Productivity
If pre-defined metrics are used for operator evaluation, then evaluation can be performed, but insight cannot be determined beyond pre-defined metrics
Solution Approach 1:
The patent transforms the evaluation approach from using fixed pre-defined metrics to using dynamic machine vibration parameters. By analyzing changes in vibration frequency, amplitude, and patterns, the system discovers new insight dimensions that were not predetermined, allowing adaptive evaluation criteria to emerge from the data
4Measurement precision
If machine learning based classification models are used for driver evaluation, then skill assessment can be performed for specific cases, but the models are not applicable to other use cases such as operators in manufacturing plant
Solution Approach 1:
The patent creates a universal operator evaluation framework that works across different machines and operations by focusing on the common principle of operator-induced vibrations. The machine learning model learns general vibration patterns associated with skill levels that can be applied to various manufacturing contexts, not just specific driving scenarios
Data Source
AI summary
A system is provided for operator profiling based on pre-installed sensor measurement. In example implementations, the system extracts a set of segmented time series data associated with a unit of operation and build models which distinguish the operators by machine learning algorithms. The system uses the models to output the evaluation score assigned to each operation, identify the key movements for skilled/non-skilled operators, and recommends appropriate actions to improve operation skill or adjust the scheduling of the operators.