Reinforcement learning-based subway station lighting and emergency system

Due to its substantial capacity, the subway has emerged as a primary mode of transportation crucial for the advancement of modern cities. However, the characteristics of subway and subway stations, such as deep underground, complex structures, and no natural lighting, have brought great difficulties to emergency rescue. In case of fire and other emergencies, it is easy to cause casualties and economic losses. Based on the basic idea of reinforcement learning, combined with computer programming, a simulation platform of subway station lighting control and emergency plan system is established, and the optimal evacuation path corresponding to three emergency danger points is obtained. Then, according to the smoke flow rules in different scenarios, the lighting control of evacuation passage is dynamically adjusted to obtain results. Experimental results show that the emergency light control controller generates corresponding evacuation plan through fire alarm information and sends it to the emergency light terminal.