To improve the level of active traffic management,a short-term traffic flow prediction model is proposed by combining phase space reconstruction(PSR)and extreme gradient boosting(XGBoost)algorithms.Firstly,the traditi...To improve the level of active traffic management,a short-term traffic flow prediction model is proposed by combining phase space reconstruction(PSR)and extreme gradient boosting(XGBoost)algorithms.Firstly,the traditional data preprocessing method is improved.The new method uses hierarchical clustering to determine the traffic flow state and fills in missing and abnormal data according to different traffic flow states.Secondly,one-dimensional data are mapped into a multidimensional data matrix through PSR,and the time series complex network is used to verify the data reconstruction effect.Finally,the multidimensional data matrix is inputted into the XGBoost model to predict future traffic flow parameters.The experimental results show that the mean square error,average absolute error,and average absolute percentage error of the prediction results of the PSR-XGBoost model are 5.399%,1.632%,and 6.278%,respectively,and the required running time is 17.35 s.Compared with mathematical-statistical models and other machine learning models,the PSR-XGBoost model has clear advantages in multiple predictive indicators,proving its feasibility and superiority in short-term traffic flow prediction.展开更多
Connected and autonomous vehicle formation(CAVF)technology is considerably important for improving transportation efficiency,optimizing traffic flow,and reduc-ing energy consumption.Despite the extensive research con-...Connected and autonomous vehicle formation(CAVF)technology is considerably important for improving transportation efficiency,optimizing traffic flow,and reduc-ing energy consumption.Despite the extensive research con-ducted on trajectory tracking control and other aspects of CAVF,the quality of the extant literature varies consider-ably,and research content remains scattered.To better pro-mote the sustainable and healthy development of the CAVF field,this paper employs the mapping knowledge domain(MKD)methodology to comprehensively review and visual-ize the current research status in this domain.Based on this review,research themes,hotspots,research challenges,and future development directions are proposed.The findings suggest that the research on CAVF can be categorized into three primary developmental stages.China and the United States are the primary countries conducting CAVF research.There is a positive correlation between economic develop-ment and the generation of scientific research outcomes.Re-search institutions are predominantly concentrated in univer-sities.The field exhibits significant interdisciplinary and inte-gration characteristics,forming key research personnel and teams.It is expected that future research will concentrate on topics such as deep learning,trajectory optimization,energy management strategy,mixed vehicle platoon,and other re-lated subjects.Research on cognition-driven intelligent for-mation decision-making mechanisms,resilience-oriented for-mation safety assurance systems,multiobjective collabora-tive formation optimization strategies,and digital twin-driven formation system validation platforms represents key future development directions.展开更多
It is interesting that despite its long-term and widespread use in China,relatively little is known about the operational characteristics of a variable approach lane(VAL)in real world.Using one month of inductive-loop...It is interesting that despite its long-term and widespread use in China,relatively little is known about the operational characteristics of a variable approach lane(VAL)in real world.Using one month of inductive-loop detector data at ten dynamic approaches(intersection approaches with dynamic lane assignment)from different intersections in Hangzhou,China,this paper presents the results of a study materializing the flow characteristics of variable approach lanes by comparing them with adjacent normal-flow lanes under various operating conditions.The effectiveness of the results was examined in a case-control analysis by integrating 12 fixed approaches(without variable lane)as benchmark.It was found that the difference or similarity of flow rate between the variable lane and the normally-flowing lane differs under a variety of traffic volume,time-of-day,modeof-operation,and overhead lane-use guidance sign(OHS)location conditions.The study also revealed that while naturally there may be a difference in the flow rates between referencing lanes at fixed approaches,the flow difference percentage(FDP)at dynamic approaches is significantly higher.展开更多
基金The National Natural Science Foundation of China (No.71771019, 71871130, 71971125)the Science and Technology Special Project of Shandong Provincial Public Security Department (No. 37000000015900920210010001,37000000015900920210012001)。
文摘To improve the level of active traffic management,a short-term traffic flow prediction model is proposed by combining phase space reconstruction(PSR)and extreme gradient boosting(XGBoost)algorithms.Firstly,the traditional data preprocessing method is improved.The new method uses hierarchical clustering to determine the traffic flow state and fills in missing and abnormal data according to different traffic flow states.Secondly,one-dimensional data are mapped into a multidimensional data matrix through PSR,and the time series complex network is used to verify the data reconstruction effect.Finally,the multidimensional data matrix is inputted into the XGBoost model to predict future traffic flow parameters.The experimental results show that the mean square error,average absolute error,and average absolute percentage error of the prediction results of the PSR-XGBoost model are 5.399%,1.632%,and 6.278%,respectively,and the required running time is 17.35 s.Compared with mathematical-statistical models and other machine learning models,the PSR-XGBoost model has clear advantages in multiple predictive indicators,proving its feasibility and superiority in short-term traffic flow prediction.
基金The National Natural Science Foundation of China (No. 52302373, 52472317)the Natural Science Foundation of Beijing (No. L231023)the Beijing Nova Program (No. 20230484443)。
文摘Connected and autonomous vehicle formation(CAVF)technology is considerably important for improving transportation efficiency,optimizing traffic flow,and reduc-ing energy consumption.Despite the extensive research con-ducted on trajectory tracking control and other aspects of CAVF,the quality of the extant literature varies consider-ably,and research content remains scattered.To better pro-mote the sustainable and healthy development of the CAVF field,this paper employs the mapping knowledge domain(MKD)methodology to comprehensively review and visual-ize the current research status in this domain.Based on this review,research themes,hotspots,research challenges,and future development directions are proposed.The findings suggest that the research on CAVF can be categorized into three primary developmental stages.China and the United States are the primary countries conducting CAVF research.There is a positive correlation between economic develop-ment and the generation of scientific research outcomes.Re-search institutions are predominantly concentrated in univer-sities.The field exhibits significant interdisciplinary and inte-gration characteristics,forming key research personnel and teams.It is expected that future research will concentrate on topics such as deep learning,trajectory optimization,energy management strategy,mixed vehicle platoon,and other re-lated subjects.Research on cognition-driven intelligent for-mation decision-making mechanisms,resilience-oriented for-mation safety assurance systems,multiobjective collabora-tive formation optimization strategies,and digital twin-driven formation system validation platforms represents key future development directions.
基金supported by“Pioneer”and“Leading Goose”R&D Program of Zhejiang Province of China(No.2022C01042)the Natural Science Foundation of Zhejiang Province of China(Nos.LGF21E080002 and LR23E080002)+1 种基金the National Natural Science Foundation of China(No.72361137006)the Alibaba-Zhejiang University Joint Research Institute of Frontier Technologies.
文摘It is interesting that despite its long-term and widespread use in China,relatively little is known about the operational characteristics of a variable approach lane(VAL)in real world.Using one month of inductive-loop detector data at ten dynamic approaches(intersection approaches with dynamic lane assignment)from different intersections in Hangzhou,China,this paper presents the results of a study materializing the flow characteristics of variable approach lanes by comparing them with adjacent normal-flow lanes under various operating conditions.The effectiveness of the results was examined in a case-control analysis by integrating 12 fixed approaches(without variable lane)as benchmark.It was found that the difference or similarity of flow rate between the variable lane and the normally-flowing lane differs under a variety of traffic volume,time-of-day,modeof-operation,and overhead lane-use guidance sign(OHS)location conditions.The study also revealed that while naturally there may be a difference in the flow rates between referencing lanes at fixed approaches,the flow difference percentage(FDP)at dynamic approaches is significantly higher.
