Highway traffic status recognition method based on bat algorithm support vector machine

Abstract

The invention relates to a bat algorithm support vector machine-based highway traffic state recognition method. The method includes the following steps that: S1, traffic state parameter data and running state data are obtained, and data sets are divided into a training set and a test set; S2, the parameters of a support vector machine are set, a bat population is constructed and initialized, an optimal bat position and a fitness value are calculated; S3, bat algorithm parameters are updated, a random number is generated for each bat individual, if rand1 is larger than R<t>i, random disturbanceis generated near an optimal solution, thus, the method shifts to local search; S4, a genetic algorithm is adopted to optimize the bat individuals; S5, a random number is generated for each bat individual, if rand2 is smaller than A<t>i, and fi is larger than f<*>, a pulse rate and loudness are updated; S6, the bats are rearranged, so that an xbest is obtained, whether a maximum number of iterations is reached is judged, and the optimal penalty parameters c and g of the support vector machine are determined; and S7, the training set is inputted into the support vector machine model so as to perform training, and an outputted predicted state is compared with the state of the test set, so that recognition accuracy can be calculated.

Description

technical field [0001] The invention relates to a highway traffic state recognition method, in particular to a highway traffic state recognition method based on a bat algorithm support vector machine. Background technique [0002] With the increase of highway traffic demand, traffic congestion, traffic accidents, tail gas pollution and other problems continue to increase, seriously endangering road traffic safety. The research on traffic status recognition enables these problems to be presented in a predictable way, providing dynamic decision-making basis for traffic participants and commanders. The research methods of traffic state recognition are roughly divided into direct method and indirect method. The early artificial traffic evacuation and the method of identifying traffic state through video image monitoring belong to the direct method; The traffic state method belongs to the indirect method. [0003] Support vector machine (Support Vector Machine, SVM) has the cha...

AI Technical Summary

Problems solved by technology

The parameter setting of the support vector machine affects its classification accuracy.

Images