A Transformer-based Method for the Single-target Location in Wireless Sensor Networks

In wireless sensor networks, it is crucial to determine where nodes are or where events occur. Obtaining nodes' accurate location information has become an essential technology of wireless sensor networks. Unlike the traditional statistical parameter estimation methods, this paper attempts to employ deep learning technology to solve the problem of single signal source location based on the angle of arrival to improve the location accuracy by practice learning. Most current algorithms use statistical optimization to solve the target point coordinates in the single target location task. These methods have the disadvantages of poor generalization ability and the inability to learn the errors in the localization environment effectively. To solve the above problems, this paper converts the solution domain of the positioning task to the image rendered by the intuitive geometric data. We then build models using computer vision and deep learning knowledge. This paper proposes the Transformer-based method to solve the target localization task.