Display Device Lifetime Prediction via Machine Learning Degradation Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing display devices face challenges in accurately predicting their lifetime due to non-uniform degradation rates of pixels, leading to issues like afterimages and color distortion, which current methods fail to address efficiently.
Innovation Solution
A method involving the creation of a machine learning model based on prior degradation rate data to estimate pixel degradation rates, using models like Linear Regression and Neural Networks, and applying these models to predict future degradation rates, thereby reducing the time required for experimental data collection and improving estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If degradation rate data is collected through actual experiments for each pixel, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent applies preliminary action by collecting degradation rate data from a first period through actual experiments and storing it in advance. This pre-collected data is then used to train a machine learning model, which can predict degradation rates for subsequent periods without requiring additional experimental measurements, thus reducing the time loss for data collection while maintaining measurement precision.
Solution Approach 2:
The patent uses machine learning to create a predictive model that copies the degradation patterns observed in the first period. The model learns from actual experimental data and generates predicted degradation rate data for future periods, effectively creating a virtual copy of the degradation process that eliminates the need for continuous physical experimentation.
2Reliability
If degradation rate data is collected through actual experiments for each pixel, then reliability is improved, but productivity decreases
Solution Approach 1:
The patent collects degradation rate data in advance during a first period and uses it to train a machine learning model before the actual lifetime prediction is needed. This preliminary data collection and model training ensure reliable predictions can be made quickly when required, improving both reliability and productivity.
Solution Approach 2:
The machine learning model creates a predictive representation of pixel degradation behavior based on historical data. This model can rapidly generate reliable degradation rate predictions for multiple pixels simultaneously, significantly improving estimation process efficiency compared to individual experimental measurements while maintaining prediction reliability.
3Productivity
If machine learning models are used to predict degradation rates, then productivity is improved, but measurement precision may be compromised
Solution Approach 1:
The patent trains the machine learning model in advance using degradation rate data collected from actual experiments during a first period. This preliminary training ensures the model learns accurate degradation patterns from real data, maintaining measurement precision while enabling fast predictions during the second period.
Solution Approach 2:
The patent uses degradation rate data from actual measurements as training feedback for the machine learning model. The model continuously learns from the relationship between applied voltages and observed degradation rates, improving its prediction accuracy over time while maintaining high productivity in the estimation process.
Data Source
AI summary
A method of predicting a lifetime of a display device according to an embodiment includes creating a machine learning model based on prior degradation rate data according to a degradation time for each of pixels, measuring a first degradation rate data for each of the pixels by inputting a voltage to each of the pixels, predicting a second degradation rate data for each of the pixels using the machine learning model, and estimating a degradation rate for each of the pixels according to a degradation time based on the first degradation rate data and the second degradation rate data.