Research On The Method And Application Of Traffic State Recognition In Key Urban Areas Based On Travel Big Data

In recent years,with the continuous increase in the number of vehicle owner ship and trip volumes,combined with the repeated prohibitions of illegal occupation of roads and random parking,which leads to more serious traffic congestion around schools,hospitals and other key areas.Key urban areas such as schools and hospitals are the main areas for residents to travel.The traffic operation status of key urban areas is closely related to the lives of residents.Traffic congestion not only aggravate the delays in travel,but also increase the incidence of traffic accidents,and seriously threaten the safe travel of primary and secondary school students and doctors and patients.However,the current traffic state evaluation in my country mainly focuses on the road network from a macro perspective,and there are few micro indicators that specifically evaluate important areas such as schools,hospitals,which is not conducive to the early adoption of traffic measures by the traffic management department.In order to facilitate traffic management departments to formulate diversion measures in advance,and the public to make reasonable arrangements for travel,there is an urgent need to identify and study the traffic operation status of schools,hospitals and other key areas.This paper uses the ideas and technologies of smart transportation processing to work from three levels of traffic collection data processing,information analysis,and information application.In terms of data processing,it mainly carries out traffic data correction and short-term prediction of traffic data in key areas.In terms of operating status recognition,it mainly conducts research on traffic status recognition in key areas.In terms of information application,it mainly carries out key area traffic information query and information service plan release.The following results were obtained:(1)Proposed traffic flow matrix filling net work model with missing weightsData filling is the basis for identification of traffic congestion.Considering that the data may be incomplete,the machine learning method is introduced.First,the existing deep learning method is optimized for historical data,and the ND model is constructed to compensate for the missing or damaged data.Then,get inspiration from image prediction and introduce spatial processing ideas to build traffic flow matrix filling net work model.Experiments show that in the case of completely random deletion,the model has better recovery performance.In the case of non-random missing,the model has a better recovery effect in the case of poor data missing.(2)Proposed traffic prediction model for key areas based on time series residual networkWith the goal of high prediction accuracy and low resource occupation,this paper establishes a short-term traffic prediction model for key areas based on time series residual networks.The term "time series residual network" is born to form the traffic prediction model of key areas based on the time series residual network,which mainly involves the central module and the independent module.The former is to carry out centralized calculation on the traffic congestion status of the surrounding road network in the predicted area,find common problems.The latter mainly refines the individual problems of each road,completes the establishment of miniaturized deep learning model,then,the time series residual network traffic prediction model is generated by model fusion.The model has higher accuracy,reaching 97% of Acc0.02 and requiring fewer parameters(3)Proposed the traffic state recognition model for key areas based on aggregate flow ratioThis paper establishes an evaluation model from a micro perspective.Firstly,it is verified that there is a close correlation between the change of the aggregated flow and the generation,development and dissipation of traffic congestion.Seconly,select and study the key congestion road sections in the area.Then,the rough set model is used to subdivide the traffic congestion interval.Finally,based on the relevant information of traffic monitoring,the cumulative mileage probability distribution model is established,the piecewise linear equation is formed,and the Key Regional Traffic Index model(KRTI)model is pointed out.Compared with the traditional index,KRTI has higher accuracy in evaluating regional traffic congestion.(4)Proposed a traffic information query algorithm based on DC-Top-k dynamicsThis paper further improves the Top-k query algorithm from dynamic partitioning.First,divide and conquer query method is used to form a better Top-k selection algorithm Dc-top-k.Secondly,the dynamic adaptive partition algorithm is introduced,the Dc-top-k algorithm is further optimized,and the intelligent partition function is added to form the dynamic adaptive partition Dc-top-k algorithm.Experimental results show that,compared with other selection algorithms,the dynamic adaptive partition Dc-top-k algorithm has outstanding performance in query effect and scalability,and is very suitable for parallel processing.The above four studies have completed the identification and application of traffic status in key urban areas from three levels of traffic collection data processing,State recognition,and information application.Proposed targeted evaluation methods from the micro level to provide a strong guarantee for subsequent more accurate analysis and research and proposed governance measures.Intelligent transportation system and users are inseparable from the maintenance of traffic information release,which is also the terminal of information service.The monitoring data of roads around schools and hospitals in this paper for 2019 were provided by Gaode and Beijing Public Security Traffic Management Bureau.This articleusing his' s traffic congestion visualization calculation method,at the same time with the help of the open platform of Gaode,the traffic congestion status of key areas is well presented.According to the traffic congestion situation of schools,hospitals and other key areas in Beijing,the traffic information release service plan is formulated to provide convenience for the government and users.