A Hidden Markov Model for Urban-Scale Traffic Estimation Using Floating Car Data.

Area of Science:

• Transportation Engineering
• Data Science
• Urban Planning

Background:

• Urban traffic congestion is a significant issue.
• Floating car data (FCD) offers low-cost, wide-coverage traffic data collection.
• Sparse probe data on arterial roads presents a major challenge for accurate traffic monitoring.

Purpose of the Study:

• To develop a novel traffic estimation model overcoming data sparseness in urban environments.
• To improve the accuracy and efficiency of traffic monitoring on arterial roads using FCD.
• To provide a practical solution for real-time traffic estimation with significant data loss.

Main Methods:

• A hidden Markov model (HMM) was employed to estimate traffic conditions as hidden states.
• Clustering and pattern mining algorithms were used to identify road segments with similar traffic characteristics.
• A multi-clustering strategy was adopted to balance clustering accuracy and coverage.

Main Results:

• The proposed HMM-based model demonstrated applicability, accuracy, and efficiency in real-world FCD experiments.
• The model proved effective for traffic estimation on urban arterials, even with over 70% missing probe data.
• The approach successfully addressed the challenge of sparse FCD on arterial roads.

Conclusions:

• The hidden Markov model (HMM) provides a robust solution for urban traffic estimation with sparse floating car data (FCD).
• The developed model is practical and efficient for real-time traffic monitoring on urban arterials.
• This research contributes to mitigating traffic congestion through improved data-driven traffic management.