Engine Control Neural Network Using Recombined Input Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing neural network models for internal combustion engines face challenges in improving calculation accuracy of output parameters due to limited types of input parameters, which restricts their expressive power and accuracy.
Innovation Solution
The implementation of a neural network model with multiple neural network units and input layers, where different combinations of input parameters are input to each unit, increasing the total number of input parameters and generating diverse intermediate state quantities, thereby enhancing the model's expressive power and accuracy without increasing the types of input parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If different combinations of input parameters are input to multiple neural network units, then calculation accuracy of output parameter is improved, but device complexity increases
Solution Approach 1:
The neural network model is segmented into multiple neural network units (first, second, third, and fourth units), each processing different combinations of input parameters through separate input layers. This segmentation allows the system to handle complex parameter relationships while maintaining manageable individual unit structures, resolving the contradiction between accuracy improvement and complexity increase.
Solution Approach 2:
The patent introduces an additional dimension by creating multiple input layers instead of using a single input layer. Each input layer processes a different combination of input parameters (e.g., first input layer processes first and second parameters, second input layer processes second and third parameters). This dimensional expansion enables the model to capture complex relationships without requiring exponential increases in individual unit complexity.
2Measurement precision
If more input parameters are used to improve calculation accuracy, then information amount for calculation increases, but the number of available input parameter types is limited
Solution Approach 1:
The limited set of input parameters is segmented and recombined across multiple input layers. Instead of requiring more parameter types, the system creates multiple combinations of existing parameters (e.g., parameters 1+2, 2+3, 3+4, 4+5 in different input layers). This segmentation approach effectively increases the information amount for calculation while working within the constraint of limited parameter types.
Solution Approach 2:
The patent transforms the problem from increasing parameter types (horizontal dimension) to increasing parameter combinations through multiple input layers (vertical dimension). By stacking input layers that process different parameter combinations, the system achieves higher calculation accuracy without needing additional parameter types, effectively moving the solution to another dimension.
Data Source
AI summary
A control device of an internal combustion engine includes a parameter acquisition unit that acquires a plurality of input parameters, a calculation unit that calculates at least one output parameter using a neural network model, and a controller that controls the internal combustion engine. The neural network model includes a plurality of neural network units and an output layer. Each of the neural network units includes one input layer and at least one intermediate layer. The neural network model inputs different combinations of input parameters selected from the input parameters to each of the input layers of the neural network units such that a total number of input parameters to be input to the neural network units is larger than the number of the input parameters.