Buslane Passenger Flow Mining Based on Travel Chain Reduction

Buslane, as the road network section with the most concentrated bus passenger flow, accurately identify bus passenger flow on buslane is a difficult and hot spot in bus route optimization and improvement research area. Based on the bus IC swipe card data, bus GPS data and bus line station data, this paper proposes an algorithm model based on bus travel chain to reduce the drop-off station and project the bus line passenger flow; combine the hidden Markov map matching algorithm, this paper builds a bus line matching road network algorithm model and a bus line multi-path calibration optimization method, which can automatically break the road network into link segment based on bus stations. Based on the bus line matching road network superimposed on the bus line passenger flow can quickly extract the bus corridor passenger flow. Applying the model results in Shenzhen show that: 2000 bus lines in Shenzhen, the bus travel chain matching rate is 92.4%, the bus line matching road network accuracy is 95%, the line network matching and single day passenger flow projection running time is about 3 hours. The proposed method of buslane passenger flow mining can efficiently and automatically complete the whole analysis process from line network matching to passenger flow extraction, and the high degree of automation and mining accuracy of buslane passenger flow mining has important application value for the optimization of city-level buslane settings.