Solving the Missing Data Problem in Urban Traffic Estimation with Principal Component Analysis

With the widespread deployment of intelligent transportation systems (ITS), there is an urgent need for observing real-time traffic conditions and trends more reliably. However, due to sensor and communication errors, the probe vehicle data commonly used in traffic estimation often face a serious missing data problem. To solve this problem, this paper proposes an algorithm based on principal component analysis (PCA). First, the missing data problem is described with a real probe car data. Then, an algorithm based on PCA is proposed to estimate the missing elements in the traffic condition matrix. Finally, some experiments were conducted to evaluate the performance of the proposed algorithm with a real probe dataset of over 10,000 vehicles in Beijing, China. Results of these experiments indicate that the proposed algorithm outperforms other competing algorithms significantly, including the KNN (K-Nearest Neighbor) algorithm and the baseline algorithm.