Support Vector Machine-Based Approach to Data-Layer Multi-Source ITS Data Fusion

Through a characteristic analysis of multi-source Intelligent Transportation System (ITS) data fusion on data layer combined with the concept of support vector machine (SVM), this article proposes a multi-source ITS data fusion technique using SVM, and designs a corresponding implementation approach from perspectives of SVM training, training result evaluation and SVM test. A comparison of data outputs derived from the SVM-based fusion and the real data before the SVM-based fusion, when the proposed technique is applied to a set of two-source traffic flow data from ShangJie on-ramp of BanShen highway in Japan, demonstrates that the proposed SVM-based fusion approach can implement the data quality control effectively, which improves the level of the data accuracy.