Method and system for predicting vehicle braking intention based on hybrid learning method

Abstract

The invention discloses a method and system for predicting vehicle braking intention based on a hybrid learning method, which belongs to the field of vehicle braking intention prediction, and can perform braking control and management of braking energy in advance, thereby improving the braking performance of the vehicle , Improve energy recovery efficiency. The method of the present invention is as follows: first collect data and perform smoothing and normalization processing on the collected data; then based on the fuzzy C-means algorithm of the unsupervised learning method, the braking intentions are clustered into four categories: slight braking, normal braking, forced braking Then, select the feature parameters based on the feature selection algorithm ReliefF; finally, based on the supervised learning algorithm Adaboost, train the braking intention prediction model offline for online prediction. The system mainly includes: a pressure sensor, a motor controller, a battery controller, a gearbox controller, a CAN bus, peripheral circuits, and an electronic control unit ECU. The prediction model obtained by offline training is located in the ECU for online prediction.

Description

technical field [0001] The invention belongs to the field of vehicle braking intention prediction, in particular to a method and system for vehicle braking intention prediction based on a hybrid learning method. Background technique [0002] The rapid development of the automobile industry, while bringing convenience to people, has also led to traffic jams, frequent traffic accidents, resource shortages, environmental pollution and other problems. Therefore, intelligent driving technology and new energy electric vehicle technology have emerged as the times require. On the one hand, the prediction of braking intention is widely used in intelligent driving assistance systems and automatic driving. It can detect the dangerous emergency conditions of the vehicle and the abnormal behavior of the driver in time, and control the safety braking in advance to ensure the safety of the vehicle system. Safety performance; on the other hand, the prediction of braking intention can be use...

AI Technical Summary

Problems solved by technology

The identification methods of the driver's braking intention can be roughly divided into two categories: one is to obtain the current state of the vehicle and the surrounding environment through the sensors installed on the vehicle system, and through the analysis of the acquired data, based on fuzzy logic, machine learning, etc. The driver’s braking intention is identified; the other is to identify the braking intention by collecting the driver’s own responses such as electroencephalogram signals (EEG) and electromyography signals (EMG). This method has high accuracy and can Realize the prediction of braking intention, but the extraction of brain signals and muscle signals is a particularly complicated process, and the required experimental equipment is also extremely expensive, so the practical value is not high at this stage

Images