A Road Network Analysis And Traffic Congestion Relief Model Based On Taxi GPS Data

Nowadays,more and more people love to travel by car with the development of the economy.However,the development of transportation infrastructure is unable to keep pace with the increase of cars,and it leads to many problems such as traffic congestion,air pollution and so on.In order to alleviate this problem,we analyze the road network of Shenzhen with the GPS data of taxi in that city.We first clean the dirty data,and show the traffic status of road segment and taxi through the visualization techniques.With the visualized picture,we find that the usage of roads and traffic regions is unbalanced.Some regions are clear while most regions of the city are under heavy traffic.In order to utilize these free regions to relief traffic congestion,we partition the city into several regions according to the distribution of taxis.Then we use Markov Chain to model the traffic,and use the equilibrium distribution to predict the status of later traffic.If the city goes to be congested,we use a method like SMO to guide some cars moving through light-traffic regions,and then the usage of regions is balanced and the traffic will not be congested.Next,we designed a distributed system to accelerate the algorithm.We have studied network IO techniques,serialization techniques,memory database,Hash algorithm and Java concurrency programming.We select the most suitable techniques to make our system efficient and safe.Finally,we process all the taxi data of 32 days and try to alleviate all the congested traffic situations.After processing,93% situations are not congested any more.We visualized the final result in our paper with detailed illustration.