01555289

Component

With the development of information technology, traffic states can be collected dynamically from floating car system with low-cost. But owing to the low penetration rate and random routes of floating cars, the traffic state of road network derived by the floating car system is usually incomplete and noised, which depresses its performance on the dynamic traffic management and related information applications. In this paper, the authors propose to estimate the traffic state of entire road network by combining the floating car system and weighted tensor completion method. Firstly, the authors form the network state into appropriate tensor model in which the observed noised entries represent the link information collected by floating car. Then the traffic state estimation problem is translated into a tensor recovery and completion problem. Leveraging the low intrinsic-dimensionality of traffic state and quality of state collected by floating car data, a weighted tensor completion method that weighting observed entries according to the quality of collected information was introduced to remove state noise and complete missing state. In the experiment, the simulated data are used to evaluate the strategy and results show that the proposed method tends to deliver high-quality solutions.

This paper was sponsored by TRB committee AHB15 Intelligent Transportation Systems.

01550057

15-5856

English

22p

2015

Transportation Research Board 94th Annual Meeting

Digital/other

Figures; References; Tables

Algorithms; Floating car data; Methodology; Tensor analysis; Traffic estimation

Data and Information Technology; Highways; I72: Traffic and Transport Planning

Transportation Research Board Annual Meeting 2015 Paper #15-5856

TRIS, TRB, ATRI

Dec 30 2014 1:57PM