Accurate and efficient map matching for challenging environments

We present the SnapNet, a system that provides accurate real-time map matching for cellular-based trajectories. Such coarse-grained trajectories introduce new challenges to map matching including (1) input locations that are far from the actual road segment (errors in the orders of kilometers), (2) back-and-forth transitions, and (3) highly sparse input data. SnapNet addresses these challenges by applying extensive preprocessing steps to remove the noisy locations and to handle the data sparseness. At the core of SnapNet is a novel incremental HMM algorithm that combines digital map hints and a number of heuristics to reduce the noise and provide real-time estimation. Evaluation of SnapNet in different cities covering more than 100km distance shows that it can achieve more than 90% accuracy under noisy coarse-grained input location estimates. This maps to over 97% and 34% enhancement in precision and recall respectively when compared to traditional HMM map matching algorithms. Moreover, SnapNet has a low latency of 1.2ms per location estimate.