Method for real-time prediction of vehicle working condition

Abstract

The invention relates to a method for the real-time prediction of a vehicle working condition. According to the method, a Markov model with an optimized prediction condition is adopted for speed prediction, and a BP neural network can compensate an error of the speed prediction; the method comprises the specific steps that the speed and acceleration of the vehicle working condition are recorded ona two-dimensional graph and gridded, grids without states are deleted, the state of the working condition is encoded into a dimension, a formula is utilized to calculate the state transition probability of adjacent moments, the current state is determined as a value, then a group of random numbers are established, the state of the next moment is predicted as another value according to the formula, the rest can be done in the same mode, the predicted speed can be found and obtained from a table, and the compensation speed is kept by using the trained BP neural network according to the actual speed. By means of the novel method for the real-time prediction of the vehicle working condition, additional hardware or sensors are not needed, and meanwhile the calculation efficiency is ensured; the vehicle working condition is predicted according to a state transition matrix established based on recorded vehicle working condition data, no hardware cost for a vehicle needs to be additionally increased, and the method has adaptability to different cycle working conditions.

Description

Technical field [0001] The invention relates to a method for real-time prediction of vehicle operating conditions. Background technique [0002] Nowadays, hybrid electric vehicles (HEV) are widely used because of their potential to improve vehicle fuel economy and reduce emissions. The performance of energy optimization management in hybrid electric vehicles (HEV) and plug-in hybrid electric vehicles (PHEV) , Both in terms of accuracy and computational efficiency, are highly dependent on the prediction of future vehicle operating conditions. [0003] At present, the academia has proposed a variety of different vehicle working condition prediction methods. According to different situations such as whether to use remote sensing or on-board sensor information, these vehicle working condition prediction methods can be divided into two categories. [0004] The operating condition prediction methods using remote sensing or on-board sensor information have different methods and algorithms....

AI Technical Summary

Problems solved by technology

[0004] The working condition prediction method using remote sensing or on-board sensor information has different methods and algorithms, however, the vehicle working condition prediction method using remote sensing or on-board sensor information will incur higher costs, and the vehicle vehicle control unit (VCU) will be more expensive. is complex and computationally more complex

Images