Vehicle driving condition prediction method based on Markov chain and neural network

Abstract

The invention discloses an automobile driving condition prediction method based on a Markov chain and a neural network, which mainly comprises the following steps: 1) obtaining characteristic parameters representing the automobile driving condition and n historically approaching vehicle speeds Vt-11, Vt-12,... Vt-1n; 2) obtain a Markov chain vehicle speed trend prediction model; 3) outputting a trend prediction result of the second set of data. 4) Principal component analysis is used to reduce the dimension of the first group of data and the third group of data. 5) obtaining a neural network ANN1 vehicle speed initial prediction model; 6) outputting initial prediction data of the second set of data. 7) obtaining a neural network ANN2 fusion model; 8) obtaining the initial forecast data ofthe third group of data and the trend forecast result of the third group of data. The invention provides a more effective prediction algorithm for the short-term working condition prediction, which isa key problem in the prediction energy management strategy of the hybrid electric vehicle.

Description

technical field [0001] The invention relates to the field of hybrid electric vehicle control, in particular to a method for predicting vehicle driving conditions based on Markov chains and neural networks. Background technique [0002] Hybrid electric vehicles have the advantages of improving environmental pollution, oil crisis, and reducing the cost of vehicle use. Therefore, at present, automobile manufacturers all over the world focus on hybrid electric vehicles as their development targets. [0003] The energy management strategy has a significant effect on improving the fuel economy of plug-in hybrid electric vehicles, but the rule-based energy management strategy cannot achieve the optimization effect, and the optimization-based energy management strategy has poor real-time performance, so there is a real-time And predictive energy management strategies for optimizing performance, such as model predictive control (MPC). MPC can coordinate the transmission system compo...

AI Technical Summary

Problems solved by technology

[0004] The Markov chain prediction model is a method for studying an uncertain system. Although the Markov chain has a tendency to predict the entire working condition, the short-term accurate prediction effect needs to be further improved to meet the needs of MPC.

[0005] Machine learning (Machine Learning) is a type of algorithm that automatically analyzes and obtains laws from data, and uses the laws to predict unknown data, but the prediction performance of neural networks is severely limited by training and over-inference.

[0006] However, a working condition prediction model that has the advantages of both self-adaption and robustness has not yet been reported. Therefore, it is necessary to develop a model with strong robustness, high prediction accuracy, not easy to fall into local optimum, and applicable to various driving conditions. Predictive models of operating conditions are very important for plug-in hybrid vehicles

Images