Reinforcement learning intelligent traffic light control method and system based on Internet-of-things platform

Abstract

The invention provides a reinforcement learning intelligent traffic light control method based on an Internet-of-things platform. The reinforcement learning intelligent traffic light control method mainly comprises the following steps of: step 1, completing simulation of urban traffic and traffic flow through microscopic traffic simulation software; step 2, acquiring equipment state information from a simulation equipment end through a OneNET cloud platform, issuing the equipment state information to a control end by the OneNET, receiving a control command, and sending the control command to the simulation equipment end to realize information communication between the simulation traffic flow and the control end; and step 3, using a DDQN neural network model to realize real-time coordination control of traffic lights according to intersection traffic flow conditions, and minimizing average waiting time of all vehicles. A reinforcement learning agent controls each intersection, each intersection trains the corresponding reinforcement learning agent, and communication is realized by sharing information among the agents. The invention further provides a reinforcement learning intelligent traffic light control system based on an Internet-of-things platform.

Description

technical field [0001] The present invention involves digital twin technology, deep reinforcement learning algorithm, and OneNET cloud platform. In particular, it involves mapping the status of physical equipment on the Internet of Things platform, using the Internet of Things platform as a transfer platform between the device end and the control end to complete information communication, and realizing the basic control function of traffic flow at multiple intersections, that is, all traffic vehicles Intersections have the shortest average wait times. Background technique [0002] In recent years, with the development of society and economy, the number of automobiles in our country has increased rapidly, resulting in traffic jams, high frequency of traffic accidents, traffic environmental pollution, traffic chaos and a series of problems, which seriously affect the current social and economic development. and people's lives. At the same time, the construction of transporta...

AI Technical Summary

Problems solved by technology

Traffic congestion has always been a key problem in designing efficient infrastructure, but traffic planning problems are difficult to solve due to the rapid growth of traffic demand

Congestion occurs when the number of vehicles attempting to use public transport routes exceeds the latter's capacity, and in worse cases leads to severe degradation in the use of available infrastructure

[0003] At present, methods such as number lottery and plate restriction widely used in major cities to relieve traffic pressure have caused a lot of inconvenience to people's travel.

Traditional fixed-logic traffic signal controllers cannot make flexible adjustments according to changes in traffic conditions, which will greatly waste personnel costs

Images