A Method for Spatiotemporal Conditioning of Vehicles Guided by Deep Reinforced Neural Networks

Abstract

The invention discloses a vehicle spatio-temporal scheduling method guided by a deep enhanced neural network. By acquiring road network information and map information, a spatio-temporal prediction model is constructed; The spatiotemporal feature vector corresponding to the vehicle information; for a certain intersection, input the spatiotemporal feature vector of the vehicle into the deep enhanced graph neural network and classify it, and obtain the probability of the vehicle going in different directions at the intersection. The present invention can more accurately analyze the dynamic changes of the road traffic network, has strong adaptability to complex and changing road conditions and can quickly find its optimal solution, and can also realize fast scheduling for relatively unfamiliar road conditions, and is suitable for complex road conditions Flow change state.

Description

technical field [0001] The invention belongs to the technical field of traffic management, in particular to a vehicle spatio-temporal temperature method guided by a deep reinforcement neural network. Background technique [0002] For domestic regions, the main means of transportation of means of production is railway or road, and the dispatching system for railway transportation is becoming more and more mature. Compared with the road network, the railway network is relatively sparse, the traffic flow in a single time period is small, the change is relatively simple, and the accident rate is much lower than that of the road. Through manual scheduling of railway partitions, supplemented by more efficient traditional algorithms, it can roughly meet the punctuality of railway transportation, and also has a certain ability to control emergencies. In terms of highways, it is impossible to directly follow the dispatching method of railways. On the one hand, compared with the rai...

AI Technical Summary

Problems solved by technology

In terms of highways, it is impossible to directly follow the dispatching method of railways

On the one hand, compared with the railway transportation system with stations as nodes, the road network is denser, with many branches and forks, making it difficult to manually dispatch

Even with the help of traditional algorithms, its excessive complexity will make the algorithm difficult to put into practical use

On the other hand, highway traffic changes are more complex, traffic jams or traffic accidents and other unexpected situations are frequent and difficult to predict in advance, and manual scheduling based on experience and algorithms based on static graph data are also incapable

From another point of view, the current manual dispatching system is mostly oriented to urban traffic, with the purpose of ensuring the smooth flow of the road network, and it is difficult to guarantee the punctuality of vehicles

Images