Study On Traffic Incident Detection And Traffic Evanesce Technologies For Freeway

This paper is supported by the projects of"Study on the key theories and techniques of freeway intelligent incidents surveillance and control system"funded by China National Natural Science under grant 50178072, and"Study on freeway automated incidents detection system"funded by Jilin Province Science and Technology Development Program under grant 20010586-1.This paper is supported by the projects of"Study on the key theories and techniques of freeway intelligent incidents surveillance and control system"funded by China National Natural Science under grant 50178072, and"Study on freeway automated incidents detection system"funded by Jilin Province Science and Technology Development Program under grant 20010586-1.Traffic incidents on freeway occur frequently as the rapid development of freeway and traffic demand. Study on the technologies of incidents detection and traffic evanescence system has been paid much attention by both domestic and overseas traffic fields, which has been the important guarantee of"safety, celerity, efficiency, coziness and environmental"on freeway, to insure safety and efficiency of freeway. By reviewing the study history and actuality of incidents detection and traffic evanescence system, this paper puts forward the current study aim and sense of freeway incidents detection and traffic evanescence technologies. Basing on summarizing the experience, achievement and problems existing, this paper analyses and evaluates the former incidents detection algorithms and control methods, and fixes the idea, contents and keystones of this paper.At first, this paper designs functions of FIDTES by analyzing demand of this system, uses advanced experience of freeway traffic incidents detection and evanescence both domestic and overseas, and then frames the freeway traffic incidents detection and evanescence united system basing on combination of fixed detector and GPS data collection techniques, making use of advanced techniques and artificial intelligence, and expatiates the functions and characteristics of each subsystem. This system includes traffic information collection subsystem, geographical information subsystem, data transmission subsystem, traffic incidents detection and alarm subsystem and traffic control and guidance subsystem. Each subsystem carries on rapid incidents detection and effective congestion guidance, according to the working process of FIDTES.The AID algorithms existed detects traffic incidents mostly basing on fixed detector data. But it indicates that these algorithms can't obtain satisfied results both in Detection Rate and in False Alarming Rate, when contrasting and analyzing several classical AID algorithms. Aiming at this disadvantage, this paper puts forward several new SVM-AID algorithms basing on fixed detector data, by analyzing the data collection and traffic flow characteristics of fixed detector. The new SVM-AID algorithms includes SVM-AID algorithm basing on fixed detector data, AID algorithm basing on MSVM, whole traffic incidents automated detection algorithm basing on MSVM-AID. In the process of putting forward the SVM-AID algorithm basing on fixed detector data, this paper firstly analyzes the characteristics of traffic flow under incidents situation, devises the traffic patterns eigenvector according to parameters chosen, and then designs several new AID algorithms aiming to different types of SVM, and uses simulation data to contrast and analyze. In the process of putting forward MSVM-AID algorithm, this paper analyzes the break characteristics of single parameter under incidents situation, and chooses multi-parameters to make up corresponding eigenvector, designs AID algorithms which combines parallel SVM and three-chasses alarming data fusion. Also, this paper puts forward the whole traffic incidents automated detection algorithm basing on MSVM according to the traffic characteristics of incidents occurrence, continuance and dissipation, which can not only detect incidents occurrence but also detect incidents dissipation. Because the three algorithms above are basing on different eigenvectors, they can detect many kinds of incidents which with different parameters changing characteristics and different detection aims. The first algorithm is suitable for all traffic parameters changing in-phase and distinctly. The second one is suitable for single traffic parameter changing distinctly. And the third one is suitable for the whole detection process. At last, this paper analyzes the validity and transplantation of the three new algorithms above. And the results indicate that the detection effects of these new algorithms are superior than classical AID algorithms.As well known, 3GS technology develops these years. Because GPS can collect instantaneous speed, travel time or travel speed of single vehicle etc., which is fixed detector can't accomplish, that increases methods and types of traffic data collection. Therefore, this paper puts forward two freeway AID algorithms basing on 3GS technology according to the congestion characteristics and single vehicle characteristics under incidents situation, aiming at the information communication mechanism of response feedback and real time feedback for incidents. The Center-Based AID algorithm basing on 3GS technology can detect all kinds of congestions FCs passing by. And the Vehicle-Based AID algorithm basing on 3GS technology can detect incidents occur on FC itself. The simulation and demonstration analysis indicate that the algorithms above can identify congestion state, congestion pattern, congestion phase, queue length, the head and the end of the queue, as well as duration time, also, they can quickly detect traffic accidents which didn't lead to congestion under low volume.Traffic evanescence is one of important mission of traffic surveillance, and study on freeway ramp control methods basing on incidents response is important way to eliminate the incidents aftereffect. This paper forms the whole ramp control pattern basing on incidents response, by analyzing capacity under incidents and characteristics of multi-ramp harmonized control. And this paper investigates method of seeking for superior value using GA, and validates the controlling effects of ramp's superior value using VISSIM4.0. The result indicate that the capacity is exerted at top level, and the system indexes of travel time, service volume, travel speed, delay and average delay have been improved remarkably, contrasting to implementing none control measures. Moreover, this paper designs several measurements of ramp meter and arterial control, and demonstrates the effects using VISSIM4.0, to fix the best FIDTES control method under different situation.At last, it is conclusions and the future work. The paper is summed up and the research process and the main contents of the thesis are looked back. At the same time, point out the innovation and the existing problems, as well as the problems that need to further study in the future.