Resource allocation method for a multi-cell communication and sensing integrated system based on DRL

Abstract

The application discloses a resource allocation method of a multi-cell communication and sensing integrated system based on DRL, and steps are as follows: (1) a deep reinforcement learning network for allocating resources to users and sensing targets is built to obtain a resource allocation action; (2) DRL and the system environment are interacted; in each interaction, all cells obtain the environment state at the current time; the system executes specific communication and sensing processes according to the resource allocation action, the current reward of the resource allocation network is obtained, and the environment state at the next time is reached; (3) a training process of DRL is performed; and (4) the trained network inputs the joint channel increase information of the users and the sensing targets as a state, dynamically adjusts a resource allocation strategy, and maximizes the sum rate of all users. The application can effectively improve the system performance, accelerate the convergence speed of the resource allocation neural network, and maximize the sum rate of all users under the condition of guaranteeing the communication quality of each user and the sensing accuracy of each target.

Images