Hybrid Vehicle Control Coefficient Prediction for Dynamic Driving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hybrid control units struggle to determine the optimal ratio of engine output to motor output in rapidly changing driving environments, leading to inefficient electrical energy and fuel consumption.
Innovation Solution
A device and method for controlling a hybrid vehicle that extracts speeds from a data set including a driving pattern and a control coefficient, learns a control coefficient prediction model using the average and standard deviation of the speeds and control coefficient, and determines the control coefficient based on the completed prediction model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If the hybrid control unit uses conventional ECMS to determine the ratio of engine output to motor output, then the control strategy is simple and easy to implement, but it cannot minimize electrical energy consumption and fuel consumption in rapidly changing driving environments
Solution Approach 1:
The patent applies dynamics by transitioning from a static ECMS control strategy to a dynamic machine learning-based control system. The LSTM neural network continuously learns from historical driving data and adapts control coefficients in real-time according to changing driving conditions, enabling the system to dynamically optimize energy consumption while maintaining adaptability to various driving environments.
Solution Approach 2:
The patent implements parameter changes by using the machine learning model to continuously adjust control coefficients based on learned patterns from historical data. The system changes key control parameters (engine torque ratio, motor torque ratio) dynamically according to predicted optimal values, allowing minimization of energy consumption across different driving scenarios rather than relying on fixed ECMS parameters.
2Ease of operation
If the hybrid control unit determines the control coefficient based on maintaining reference SOC without considering driving patterns, then the control logic is simple, but it fails to optimize energy consumption in varying driving conditions
Solution Approach 1:
The patent applies preliminary action by pre-training the machine learning model with extensive historical driving data before actual operation. The LSTM network learns optimal control patterns in advance from the training dataset, enabling it to make intelligent control decisions during real-time operation without complex real-time calculations, thus maintaining ease of operation while optimizing energy consumption.
Solution Approach 2:
The system implements self-service through the machine learning model's ability to autonomously determine optimal control coefficients based on learned driving patterns. The model automatically adjusts engine and motor torque ratios without requiring complex real-time intervention or simplified control logic, enabling the system to self-optimize energy consumption while maintaining operational simplicity.
Data Source
AI summary
Disclosed are a device for controlling a hybrid vehicle and a method thereof. The device includes a communication device that receives a plurality of data sets including a driving pattern and a control coefficient, and a controller that extracts speeds from the driving pattern, learns a control coefficient prediction model by using an average and a standard deviation of the speeds, and determines a control coefficient of the hybrid vehicle based on the control coefficient prediction model for which the learning is completed.