Road surface roughness prediction method based on random forest

Abstract

The invention discloses a road surface roughness prediction method based on random forest. The road surface roughness prediction method based on the random forest comprises the following steps that S1, for each asphalt road surface, values of a road surface roughness index IRI and road surface parameters are collected, and asphalt road surface sample data are obtained; S2, each road surface parameter is normalized to obtain a training set; S3, the training set is trained through a random forest algorithm, the road surface roughness indexes IRI are used as the output of the random forest algorithm, the road surface data are used as the input, and a trained asphalt road surface roughness prediction model is obtained; S4, road surface parameter values of a to-be-measured asphalt road surfaceare collected, and to-be-measured asphalt road surface data are obtained; S5, normalization operation is performed; and S6, the to-be-measured normalized asphalt road surface data are processed by using a trained asphalt road surface roughness prediction model, and the road surface roughness of the to-be-measured asphalt road surface is obtained. According to the road surface roughness predictionmethod based on the random forest, the road surface roughness prediction model based on the random forest is established, and the prediction precision of the road surface roughness can be greatly improved through the model.

Description

technical field [0001] The invention belongs to the technical field of road surface monitoring, and in particular relates to a method for predicting road surface roughness based on random forests. Background technique [0002] Road surface roughness (Road Surface Roughness) refers to the deviation value of the longitudinal unevenness of the road surface. Road surface roughness is an important index in road surface evaluation and road construction acceptance. It is related to the safety and comfort of driving and the impact on the road surface. Force size and service life. Therefore, the efficiency and accuracy of the measurement of pavement roughness is very important for pavement evaluation. [0003] There are three traditional methods for measuring road surface roughness: fixed-length ruler method, cross-section drawing method, and bump accumulation method, but these traditional methods are not very ideal in terms of efficiency and accuracy. At present, neural network met...

AI Technical Summary

Problems solved by technology

[0003] There are three traditional methods for measuring road surface roughness: fixed-length ruler method, cross-section drawing method, and bump accumulation method, but these traditional methods are not very ideal in terms of efficiency and accuracy. At present, neural network methods are used to predict road surface roughness It has become the main detection method. This method solves the problem of accuracy to a certain extent, but there are still two problems in the following aspects: First, there is no unified and complete theoretical guidance for the selection of neural network structure. It can only be selected by experience. If the network structure is too large, the training efficiency will not be high, and overfitting may occur, resulting in low network performance and reduced fault tolerance. If the network structure is too small, the network may not converge.

In addition, no matter how simple the problem is, the method of neural network needs to go through hundreds of times of learning, which leads to very slow prediction speed.

Images