Driving behavior safety evaluation method based on driving fingerprint

Abstract

The invention discloses a driving behavior safety evaluation method based on driving fingerprint. The method comprises the steps of S1, collecting driving behavior data; S2, calculating a characteristic index representing individual differential driving behavior according to driving behavior data; S3, selecting a training set, inputting that characteristic index into a machine learning classify for training, and calculating a driving fingerprint through the machine learning classifier; S4, taking the driving behavior data to be evaluated as a test set, and comparing the difference degree between the driving behavior data and the driving fingerprint under the normal driving state; S5, calculating the weight of the importance degree of each parameter in the driving fingerprint to the drivingsafety; S6, the parameters of the difference degree were calculated by the principal component analysis method, and the safety of driving behavior was evaluated by the factor scores. The method of the invention can fully consider the influence of individual difference factors and obtain accurate, comprehensive and reliable driver safety evaluation results.

Description

technical field [0001] The invention relates to the field of driving safety evaluation, in particular to a driving behavior safety evaluation method based on driving fingerprints. Background technique [0002] In recent years, with the development of on-board sensor technology, research in the field of traffic safety has gradually deepened. Many car manufacturers are actively developing personalized vehicle assisted driving systems to improve the overall reliability and adaptability of the system. The basis of such systems is the identification and evaluation of the individual driving characteristics of the driver. At the same time, through research on individual differences in driving behavior in recent years, it has been found that personalized driving behavior models (driving fatigue detection model, lane keeping model, and vehicle speed control prediction model, etc.) constructed with reasonable consideration of individual differences are compared with traditional models...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is aimed at the low accuracy and poor reliability of the evaluation results caused by the driver's individual differences in the existing driving safety evaluation system in the Internet of Vehicles environment, and the traditional evaluation index method cannot adapt to the personal characteristics of each driver to be evaluated. The characteristics of driving habits lead to unconvincing and fair evaluation results

Images