Microscopic urban traffic online simulation system and microscopic urban traffic online simulation method

Abstract

The invention relates to a microcosmic urban traffic online simulation system and a microcosmic urban traffic online simulation method. According to the invention, pre-training is carried out; the relationship between the simulation flow parameter and the road section speed is learned by adopting a meta-learning method; a simulation flow parameter calibration value can be obtained by only using small sample real speed data to adjust parameters of the trained basic model, so that online simulation is realized, the basic model has the advantages of few sample requirements, good generalization performance and the like; the problem that a conventional traffic simulation model cannot perform flow parameter calibration by using small samples is solved; in order to realize an optimal model on thebasis of short time, a double-layer distributed structure is adopted, and distributed structures are adopted in a distributed model parameter learning module and a distributed simulation module, so that the time spent on calculation is greatly reduced, and guidance can be provided for real-time traffic analysis requirements.

Description

technical field [0001] The invention relates to the technical field of traffic simulation, in particular to a microscopic urban traffic online simulation system and method. Background technique [0002] With the rapid development of cities and the continuous increase of car ownership, traffic congestion is becoming more and more normal, and traffic management and control are facing major challenges. The analysis and evaluation of traditional traffic control mainly rely on manual observation. The analysis ability of manual observation is very limited. It is often difficult to trace the source of the problem of monitoring dead spots, and it is impossible to quantitatively analyze it. With the increase of informatization means, the amount of data such as microwave detection data, monitoring bayonet data and geomagnetic information has shown a sharp increase. On the basis of these traffic big data, the analysis and evaluation of traffic control is gradually refined. However, si...

AI Technical Summary

Problems solved by technology

[0005] The present invention overcomes above-mentioned deficiencies, and the purpose is to provide a kind of microcosmic urban traffic online simulation system and method, the present invention adopts meta-learning method to learn the relationship between simulated traffic parameters and road section speed through pre-training, after the training is completed The basic model of the basic model can obtain the calibration value of the simulated traffic parameters only by adjusting the parameters of the real speed data of a small sample, and then realize the online simulation. The problem of sample flow parameter calibration is to realize the optimized model on a short-term basis. At the same time, the present invention adopts a double-layer distributed structure. Both the model parameter learning module and the simulation module adopt a distributed structure, which greatly reduces the Calculates time spent, which can provide guidance for real-time traffic analysis needs

Images