Research And Design Of Road Condition Perception System Based On Intelligent Termina

In recent years,China’s investment in infrastructure construction is increasing,and the mileage of new roads continues to grow.With the new roads put into use,the length of roads that need inspection and maintenance has increased dramatically.At present,the road condition detection equipment is inconvenient to carry,expensive and high operation threshold all increase the difficulty of the road administration department staff.Road inspection and maintenance are not timely,leading to the increase of bad road conditions,which will pose a certain threat to the life and property safety of the drivers and passengers.Therefore,how to improve the work efficiency of the road administration department and reduce the cost of road condition inspection equipment has become an important subject of the highway maintenance department.This thesis mainly uses the acceleration sensor and GPS module integrated by the intelligent terminal carried by the driver to conduct road condition data acquisition,Through the data processing and feature extraction,Proposed based on Smote Enn(synthetic minority oversampling technique edited nearest neighbor,Smote Enn)and Light GBM(light gradient boosting machine,Light GBM)The road condition sensing algorithm,Used to identify five road conditions: smooth,potholes,manhole cover,speed bump and railway crossing.The algorithm using Smote Enn algorithm to solve the data imbalance leads to the number of road type classification accuracy affects the whole model classification accuracy problem,using Light GBM model for road data training and classification solved the commonly used machine learning classification algorithm in the process of resource consumption,the problem of long operation time.Based on the proposed algorithm,the road condition awareness system is designed and realized,which includes road collection data app(Road-Collection-Data APP,RCD-APP)and road conditions background management system.In this thesis,RCD-APP is used for field data acquisition and application experiments to verify the feasibility of the method and the reliability of the system.After testing,the road condition perception algorithm has high accuracy,the road condition perception system runs stably,the road condition data collection,road condition view and road maintenance management functions are simple to operate,and can be widely used in the road condition maintenance and management business.After further perfecting this system,it has good engineering application value and commercial promotion value.