Data Fusion for Traffic Incident Detector Using D-S Evidence Theory with Probabilistic SVMs

摘要: Accurate Incident detection is one of the important components in Intelligent Transportation Systems. It identifies traffic abnormality based on input signals obtained from different type flow sensors. To date, development Systems has urged researchers incident area to explore new techniques with high adaptability changing site characteristics. From viewpoint evidence theory, information each sensor can be considered as a piece evidence, and such, multisensor detector viewed problem fusion. This paper proposes technique for detection, which combines multiple multi-class probability support vector machines (MPSVM) using D-S theory. We present preliminary review theory explain how multi-sensor framed context this terms incidents frame discernment, mass functions designed by mapping outputs standard into posterior learned sigmoid function. The experiment results suggest that MPSVM better adaptive classifier environment.