Research And Implementation Of Dynamic Traffic Guidance Systembased On Floating Car Data

The system of dynamic traffic guidance possesses the capability to enhance theefficiency of the vehicle travel and the use of roads and to ease the traffic congestion.Most of the traditional system of traffic guidance are based on the construction ofstatic traffic information without combining the dynamic traffic information, thus ithas less practicality. In order to solve this problem, a system of dynamic trafficguidance was invented based on floating car data, and many key technologies,theautomatic generation of dynamic road network in system, estimation model ofresistant value of the dynamic road and the algorithm of dynamic trafficguidance,were further studied. The main works are as follows:1、Based on relational data, an road network automatically generated algorithmwas proposed. To solve the high redundancy and low efficiency of the guidancealgorithm, the data for GIS was re-analyzed, the redundant data was eliminated, andthe result was stored with relational database, based on the spatial data of GIS. Thesearching space was dynamically constructed according to the input of the user, whichimproves the generation efficiency of the road network.2、The processing algorithm of floating car data was studied. To solve the errorand the lack of attribute information of the traditional floating car data, the principleof data cleaning to filter the wrong data and revised the error data of the floating carwas designed, by using the ST map-matching algorithm.3、Based on the delay at intersection and spatial topological relation, an dynamicestimation model of road resistant was constructed. It was the central part of thetraffic guidance system. With the delayed time at intersection through the floating cardata and the information of road topology, road resistant was dynamically estimated.4、An improving ant colony algorithm for traffic guidance was proposed. Due tothe low efficiency of ant colony algorithm in the large-scale network, three improvingregulation were provided, which does not only overcome the low efficiency but also more adapts to the discrete and correlated character of the floating car data.The system was integrated based on the algorithm mentioned above, and testedby using the data of6000taxis operating in60days. The result shows that the systemis stable and highly practical. The matching accuracy is93%. The dynamic trafficguidance of path produced by this system saves more than20%of the time compareto the path produced by the static information.Therefore it improves the travelefficiency.