Signal intersection adaptive control method based on improved deep Q network

The invention relates to a signal intersection adaptive control method based on an improved depth Q network, and the method comprises the following steps: obtaining vehicle passing data in an intersection environment, and calculating the real-time queuing information of vehicles; an intersection signal control agent is defined based on the queuing information and comprises an action space, a state space and a reward function, the action space is provided with a minimum green light duration constraint, the state space is constructed by using a maximum queuing principle, and the reward function is constructed by predicting the number of queuing vehicles; learning and storing an interaction strategy of the intelligent agent and an intersection environment on line based on a deep Q network; and selecting an optimal behavior strategy of the intelligent agent from the interaction strategies and executing the optimal behavior strategy so as to realize traffic signal control. Compared with the prior art, the method h