Study On Detection Algorithm Of Traffic Incidents And Detectors Distribution Of Freeway

The development of freeway compared with common highways has acquiredgreat success, which produces great economic and social benefits and has a distinctimprovement in service level. But the traffic problems in operation, such ascongestion, safety and so on, have severe impact on the whole traffic capacity andoperation efficiency, which violates the original intention of freeway construction.At the same time, with the step-by-step development of science and technology,people are gradually convinced that in order to solve the traffic problems all kindsof advanced technologies are needed, which can integrate information technology,data communication technology, automation technology and informationprocessing technology efficiently. And all these technologies will be exerted on thewhole traffic management system to set up a real-time, exact and high efficiencyintegration transportation intelligence control and management system.The main task of traffic incidents detection system is to detect and to disposeof traffic incidents on the freeway in time, so that to reduce personnel casualtiesand property loss and to avoid the occurrence of following incidents. Nowadays thescholars of other countries have developed many algorithms,this dissertation sumsup these classical traffic incidents detection algorithms and indicates that thesealgorithms are mostly based on other countries freeways'detection data and theirmain limitation is that the algorithms are lack of replantable and auto-adaptable,while our freeway has great differences with other countries in traffic flow, trafficmakeup and road establishment and so on. If these algorithms are used in ourcountry's incidents detection directly, the effect won't be good. And because thebasic principle of highway incidents detection is to identify traffic state of givensections by the analysis of the detectors data change rules, the detectors data'squickness, precision and real reflection traffic state have direct impact on thevalidity of incidents detection algorithm, and their cost benefit is the main reasonof their employ values. Therefore, it is necessary to optimize detectors distributionwhen it meets the need of traffic data precision.This dissertation aims at traffic congestion incidents and analyses theirformation, features, and measurement parameters etc. And according to theappraisal indexes these common algorithms are contrasted and analyzed. Then thevelocity, traffic volume and occupancy data of loop detectors in some domesticfreeway are analyzed and Stat. by use of traffic flow basic theories. At the sametime the traffic running character in congestion state is integrated. Then incrementidentification rule of traffic congestion appearance and disappearance in our trafficflow circumstance is put forward. By theory induction and Stat. analysis of twoimportant indexes, traffic flow and occupancy, appropriate congestion detectionindexes are established. And then the traffic states are divided according to theindexes to detect and traffic incidents are identified automatically. This algorithm istestified by CORSIM, which is micro simulation software. And the analysis resultof simulation data shows that the increment identification rule not only can reflectin time the course of traffic congestion appearance and disappearance, but also cancorrectly identify traffic congestion incidents. This identification and detection iseasy to realize, and so this algorithm has rather employment value.The common detectors distribution principles are put forward from the pointof experience and quantitative analysis aiming at incidents detection. Then theimpact of detectors density on travel time is analyzed concretely. From the point oftravel time computation, detectors distribution principle is advanced and simulationexperimentations are made to testify this principle. Finally the relation of detectorsnumber, investment and detection error is analyzed elementarily, which guidesdetectors distribution of different purposes.The research of freeway traffic incidents detection algorithm and detectorsdistribution touches upon traffic engineering, autocontrol theory, electronictechnology, information technology, internet technology and data communicationtransmission technology. In the research work, management, communication,...