A Hybrid Model for Speed Estimation Based on Single-Loop Data

The accurate, real-time traffic speed is important for a successful freeway traffic management system. Unfortunately, vehicle speeds cannot be directly measured by single-loop detectors. In this paper, a hybrid model of single-loop data for speed estimation is presented, different models are used to estimate the speed according to different traffic conditions such as free flow and congested flow. As for the coefficient of the model, it can be estimated by using the known single-loop outputs. The whole algorithm is simulated and implemented by Matlab simulation software. Observed from the results, speeds estimated by this method are very close to the speeds observed by the corresponding dual-loop detector. Compared to the commonly adopted speed estimation algorithm, the proposed method improves estimation accuracy significantly.