Placement of Access Points for Indoor Positioning based on DDPG

With the development of the Internet of Things, location-based services are becoming ubiquitous. Rapid and accurate deployment of access points (AP) in indoor environments is an effective way to improve the positioning accuracy. In this paper, a method is proposed for optimal deployment of indoor positioning access points based on Deep Deterministic Policy Gradient Algorithms (DDPG). With this method, the APs, in their initial states, are randomly placed in the target area under the premise of ensuring full network coverage. Each AP is regarded as an agent, and the optimal objective of AP deployment is defined as achieving the maximum Euclidean distance of the reference signal. The priority experience replay mechanism is introduced in this process, which guides the behavior by performing a series of actions and interacting with the environment to obtain the maximum reward for the agent. Simulation experiments are carried out to evaluate the performance of the proposed method. The results show that the proposed deployment method can converge quickly. Compared with the random deployment method and the maximization-minimization method, the proposed deployment method can effectively improve the positioning accuracy.