Research On User-oriented Data Quality Management Of Pavement Condition

Pavement condition data refers to various kinds of data that can represent the damage condition and road performance of the road surface,which is mainly used to predict the road performance and life and provide decision to manage pavement.With the increase of road mileage,service life and the improvement of detection technology,the volume of pavement condition data increases exponentially.Data quality issues inevitably occur in the process of data management,leading to unreliable data analysis,inaccurate decision-making,loss of a large number of resources and other problems.Therefore,it is very important to manage pavement condition data quality.This paper aims to study effective and efficient methods of pavement condition data quality management,and does research and draws conclusions from the following four aspects:(1)Based on road data forms and characteristics,through the literature research method,after the merit and demerit of the quality management and data quality management methods are compared,the most suitable for road data quality management method,Total Data Quality Management,is put forward.By the method,pavement condition data characteristics are defined in the basic level which means the raw data are extracted from pavement directly and advanced level which refer to index data included specific meaning.(2)Based on the analysis of user demand,the users of pavement condition data are divided into two categories: road managers and road users,which are sequentially subdivided according to their functions,meanwhile their interest points are analyzed.Combined with the Total Data Quality Management method,four main indexes affecting the data quality of pavement condition are put forward according to user requirements,namely,repeatability,resolution,accuracy and availability.(3)Based on the integrated pavement performance evaluation car,in terms of a problem which is in the actual detection deviation of IRI measurement will be caused by different testing conditions,through the experiment,the relationship between independent variables relative to the conditions and the deviation is analyzed.Thus,a optimization method based on partial least squares is put forward,and after experimental verification,by using the optimization method the measurement error is reduced from 20% to about 5%.(4)In terms of the phenomenon that there are non-uniform standards for automatic recognition of distress in road detection,this paper takes crack identification as an example and gives suggestions on the definition methods of four quality indicators,including precision,recall,accuracy and repeatability,in combination with user requirements and detection technology merit.In the framework of Total Data Quality Management,from the pavement condition data definition,data quality index definition,and data quality analysis to data quality improvement methods,the pavement condition data quality was managed step by step.In conclusion,there are certain theoretical significance and practical value in this paper for the area of road management,data quality improvement,data quality standard definition,and so on.