Currently, there are a variety of traffic flow data detection devices. For the limit of the measurement precision of the devices, the traffic flow data from one kind of device is usually inaccurate. The fusion of the data from multiple kinds of devices can improve the accuracy. A data fusion algorithm based on the multi-attribute decision is proposed, with all kinds of detector data as the decision scheme, measurement precision, deviation to the historical average, and deviation to the average as decision attributes, which establishes a decision matrix to determine optimal fusion value. Meanwhile, a traffic state identification method based on fuzzy inference is proposed, which constructs the fuzzy relationship among flow rate, occupancy, and traffic state to estimate traffic state. The experiment shows that, through the application of the data fusion algorithm and the traffic state identification method, the detection accuracy of traffic flow in the test sites is increased by 8.1% and the estimation results of the road traffic state is very consistent with the actual situation.
展开▼
