CNN Matrix Layout for Battery Deterioration Prediction Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating the deterioration state of devices like power storage elements and vehicle batteries using convolutional neural networks do not effectively utilize peripheral data that indirectly affects performance and deterioration, leading to suboptimal prediction accuracy.
Innovation Solution
A method for generating matrix data for convolutional neural networks that incorporates both primary and secondary data influencing temporal changes, where primary data with high influence is alternately arranged with secondary data of lower influence using a kernel or filter, allowing for more comprehensive convolution operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only primary data with high influence is used for convolution operation, then the model complexity is reduced, but the prediction accuracy deteriorates because peripheral data that indirectly affects performance is not utilized
Solution Approach 1:
The patent segments data into two categories: primary data (high influence) and secondary data (peripheral data with indirect influence). This segmentation allows the system to process both types of data systematically, ensuring that neither important direct factors nor useful indirect factors are omitted, thereby improving prediction accuracy while maintaining manageable complexity
Solution Approach 2:
The patent merges primary data and secondary data into a unified data structure that feeds into the convolutional neural network. By combining these different types of data, the model can simultaneously consider both direct and indirect影响因素, achieving more comprehensive and accurate predictions without excessive complexity
2Measurement precision
If peripheral data with lower influence is excluded from input, then the data processing speed is improved, but the estimation accuracy deteriorates due to loss of relevant information
Solution Approach 1:
The patent applies local quality by treating different data types with different levels of importance. Primary data receives higher weight and attention in the processing pipeline, while secondary peripheral data is processed with appropriate but lower priority. This differentiated approach ensures accurate estimation by capturing all relevant information while optimizing processing efficiency through selective attention
3Loss of information
If all available data including peripheral data is processed, then the information completeness is improved, but the computational cost increases
Solution Approach 1:
The patent extracts and processes only the most relevant features from both primary and secondary data through the convolutional neural network's feature extraction capability. Rather than processing all raw data equally, the system selectively extracts meaningful patterns and relationships, achieving information completeness while reducing unnecessary computational energy consumption
Data Source
AI summary
In a method for generating matrix data for a convolutional neural network that performs a convolution operation based on the matrix data in which vehicle information is arranged as an matrix element, the matrix data is composed of predetermined time-series data in which each row changes continuously in terms of time in an arrangement direction of each column, the time-series data is composed of first data of which degree of influence on the convolution operation is high and second data of which degree of influence is lower than the first data, the convolution operation is performed using a kernel that partitions the matrix data into the rows and columns corresponding to a predetermined coefficient, and at least one row of the first data is arranged for each set of rows corresponding to the coefficient.