High-speed train parking control method and system based on deep learning

Abstract

The invention discloses a high-speed train parking control method and system based on deep learning, and relates to the technical field of rail transit management and control. The method comprises thesteps of: obtaining a training data set; establishing a convolutional neural network; training and optimizing the convolutional neural network by using the training data set to obtain an optimized convolutional neural network; acquiring actual operation data of a to-be-controlled train; inputting the actual operation data into the optimized convolutional neural network to obtain a parking position of the to-be-controlled train; judging whether the parking position of the to-be-controlled train is 0 or not; if so, outputting a brake command; and if not, returning to the step of acquiring the actual operation data of the to-be-controlled train. Accurate parking of the high-speed train can be achieved without a large amount of cost.

Description

technical field [0001] The invention relates to the technical field of rail traffic management and control, in particular to a high-speed train parking control method and system based on deep learning. Background technique [0002] Rail transportation has been fully developed in my country due to its advantages of energy saving, comfort, environmental protection, safety, punctuality and convenience. High-speed railways can not only greatly reduce the travel cost of passengers and reduce the burden on passengers, but also drive the economy along the line and narrow regional differences. As of the end of 2019, the operating mileage of China's high-speed railways has reached 35,000 kilometers, accounting for more than two-thirds of the world. In 2019, the number of passengers sent by China's high-speed railway sports trains ranked first in the world, with a cumulative total of 2.29 billion for the year person-times. As one of the seven major areas of new infrastructure, high-...

AI Technical Summary

Problems solved by technology

[0012] The above model is an accurate mathematical model that comprehensively considers all factors. Since it contains many nonlinear characteristics (such as air resistance, etc.) and the time-delay characteristics of the train, it is necessary to establish a very complex motion model, and some of the parameters are easily disturbed by the external environment. However, it is difficult to realize the change in practical application.

At the same time, different vehicles have different adaptability to the environment, and their sensitivity to parameters is also slightly different. In practical applications, the model needs to undergo a large number of field tests, which increases the cost of research and development.

[0013] In summary, the train braking system is a very complex nonlinear time-delay system, which contains a large number of parameters. The existing train braking model parameters are fixed, and the external environment (such as wind, frost, rain, snow, temperature) and the impact of train operation loss In this way, the train braking model is inaccurate, and the parking error cannot be accurately estimated. Secondly, the existing train braking model needs to go through a large number of field tests when debugging parameters, and the cost is relatively high.

Therefore, the existing models cannot dynamically perceive the impact of changes in the external environment on the train, reflect the real running state of the train, and thus cannot accurately predict the parking position of the train. If the existing model is used to accurately predict the parking position of the train to achieve precise parking of the high-speed train , a large number of field tests and parameter adjustments are required, and the research and development costs paid are huge.

Images