Xuran Ivan Li’s research while affiliated with The Hong Kong Polytechnic University and other places

Short-term traffic flow data is characterized by rapid and dramatic fluctuations. It reflects the nature of the frequent congestion in the lane, which shows a strong nonlinear feature. Traffic state estimation based on the data gained by electronic sensors is critical for much intelligent traffic management and the traffic control. In this paper, a solution to freeway traffic estimation in Beijing is proposed using a particle filter, based on macroscopic traffic flow model, which estimates both traffic density and speed. Particle filter is a nonlinear prediction method, which has obvious advantages for traffic flows prediction. However, with the increase of sampling period, the volatility of the traffic state curve will be much dramatic. Therefore, the prediction accuracy will be affected and difficulty of forecasting is raised. In this paper, particle filter model is applied to estimate the short-term traffic flow. Numerical study is conducted based on the Beijing freeway data with the sampling period of 2 min. The relatively high accuracy of the results indicates the superiority of the proposed model.

This paper proposes a novel technique to forecast day-ahead electricity prices based on Panel Cointegration (PC). The current researches on the electricity price forecasting focus on the analysis of unstable economic time series. However, due to the difference of the allocation of power resource and consumption in different regions, the time series of electricity consumption and sales price in a single region cannot contain all information in different regions. In view of the disadvantages of the time series, the panel data is introduced to investigate the long-term equilibrium relationship and the short-term adjustment relationship between electricity consumption and sales price in the paper. The fundamental and novel contribution of the paper is to apply the PC to forecast accurately in day-ahead energy market price. The whole forecasting framework shows the use of PC model in predicting price behavior. Results from the electricity market of PJM in year 2008 are reported.