Traffic state estimation device and method based on data fusion

Abstract

The invention discloses a traffic state estimation device and method based on data fusion. The traffic state estimation device comprises a pattern classifying unit, a model selecting unit and a data fusing unit, wherein the pattern classifying unit is used for classifying traffic data of one or more roads into one of a plurality of preassigned patterns; the model selecting unit is used for selecting a corresponding neural network model from a plurality of neural network models according to the classified patterns; and the data fusing unit is used for inputting traffic data into the selected neural network model for carrying out data fusion and estimating a traffic condition of one or more roads. According to the invention, a multi-pattern classifying and multi-pattern modeling mechanism is introduced, aiming at each road, different patterns are respectively independently modeled, thus the precision of the single model can be improved, and different data source types of the traffic data, different characteristics and different factors of influencing the traffic conditions are fully utilized to obtain a more realistic and accurate fusion result and traffic condition estimation.

Description

technical field [0001] The invention relates to the field of intelligent transportation systems, in particular to a data fusion-based traffic state estimation device and method. Background technique [0002] With the development of my country's urban intelligent transportation system in recent years, many important cities have built various traffic detection methods to realize real-time monitoring of urban transportation systems. These detection means include fixed detection equipment (such as microwave detectors, coil detectors) and mobile detection equipment (such as floating vehicles). These detection devices provide different traffic operation information content, but due to the different traffic detection methods, there are certain differences in the traffic information reflected by different data in terms of information form, integrity, accuracy, and consistency. It is difficult to obtain traffic information with large coverage and high accuracy only by relying on the...

AI Technical Summary

Problems solved by technology

[0006] Although the method mentioned in Patent Document 1 can integrate multiple traffic data and obtain more accurate road conditions, it has the following disadvantages: First, the method in Patent Document 1 only builds a single model for fusion, and cannot reflect different data sources There are differences in the information form, completeness, accuracy, consistency, etc. of traffic data, and it is impossible to distinguish different combinations of traffic data and traffic conditions; second, the construction of Bayesian network needs to be based on the data For the determination of the probability distribution of each node in the Bayesian network, Patent Document 1 adopts the method of actual measurement and expert verification, so long-term observation of a certain road section is required to determine the probability distribution from weather to traffic conditions. The conditional probability distribution of the conditional probability distribution makes the method more complicated and the workload is larger; the third is that in the Bayesian network fusion, only the point speed detected by the fixed detector is used, and the traffic flow and road occupancy are ignored. The point speed is the vehicle passing through The instantaneous speed at the moment when the detector is fixed is different from the average speed of the road, and the traffic flow and road occupancy rate can also reflect the traffic situation to a certain extent, and these two data should not be discarded; the fourth is in Bayesian The final result of network integration is the degree of road congestion, rather than specific speed values, which cannot provide deeper applications

Images