Travel Time Estimation from Sparse Floating Car Data with Consistent Path Inference: A Fixed Point Approach
(English)Manuscript (preprint) (Other academic)
An important application of sparse floating car data (FCD) is the estimation of network link travel times, which requires pre-processing by map-matching and path inference filters. Path inference, in general, requires some a priori assumption about link travel times to infer paths that are reasonable and temporally consistent with observations. Path inference and travel time estimation is thus a joint problem. This paper proposes a fixed point approach to the travel time estimation problem with consistent path inference.The methodology is applied in a case study to estimate travel times from taxi FCD in Stockholm, Sweden. In this case study, existing methods for path inference and travel time estimation, without any particular assumptions about path choice models or travel time distributions, are used. The results show that standard fixed point iterations converge quickly to a solution where input and output travel times are consistent. The solution is robust under different initial travel times and data sizes. Using historical initial travel times reduces bias. The results highlight the value of the fixed point estimation process, in particular for accurate path finding and route optimization.
floating car data, fixed point problem, travel time estimation, path inference, iterative process
Transport Systems and Logistics
Research subject Transport Science
IdentifiersURN: urn:nbn:se:kth:diva-167731OAI: oai:DiVA.org:kth-167731DiVA: diva2:813422
QS 20152015-05-222015-05-222015-05-25Bibliographically approved