Abstract: Since the actual transportation system features in complex time-varying changes and uncertainty of traffic flow changes，the wavelet analysis theory is applied to denoise the original traffic data，which makes the data after denoising is better to reflect the nature of traffic flow and its changing rule. Information fusion technology is able to do integrated treatment of different sensor data，to remove redundancy，to overcome the ambiguity，so it can obtain more comprehensive， more accurate and reliable information than any single data source do. Considering the accuracy and reliability of real-time road traffic prediction，the information fusion prediction method based on wavelet denoising and optimum weight is proposed， using the thought of integrating denoising technology based on wavelet analysis and modern information; furthermore， the experimental results are verified by case studies. Empirical results of the analysis show that the method proposed can improve the accuracy of traffic flow prediction effectively.

Key words: short-time traffic flow prediction, information fusion, optimal weight, wavelet analysis

摘要： 针对实际交通系统时变复杂的特征和交通流变化的不确定性，应用小波分析理论，对原始交通数据进行了消 噪处理，使消噪后的数据更能反映交通流的本质及变化规律。信息融合技术可以对不同传感器数据进行综合处 理，去除冗余、克服歧义，得到比任何单个数据源更全面、更准确、更可靠的信息。综合考虑道路交通量预测的实 时性、准确性和可靠性，运用基于小波分析的去噪和现代信息融合思想，提出了一种基于小波去噪和最优权重的 信息融合预测方法，并用实际数据进行验证。实证分析的结果表明，该方法能够有效提高交通量的预测精度。

关键词: 交通量预测, 信息融合, 最优权重, 小波分析