| In recent years,new location-based services such as global positioning system,industrial Internet of Things,automatic driving,indoor map,shared bicycle and so on have emerged continuously.High-precision positioning is considered as one of the essential ubiquitous services of mobile communication systems since 4G.However,most current positioning technologies work in lower frequency bands,and their spectrum resources are very scarce,making it difficult to meet the highprecision performance requirements of wireless positioning technology.Due to the wireless propagation characteristics of short wave length and wide frequency band in the mm-wave frequency band,which helps to achieve higher positioning signal measurement accuracy,it has naturally received great attention.However,at the same time,due to the faster signal attenuation and shorter transmission distance in the millimeter wave frequency band,more base stations need to be deployed to achieve reliable communication.In addition,5G networks are mostly heterogeneous and coexist,with numerous devices using different RATs.The proliferation of communication equipment types and quantities has led to greater random variability in channel conditions in millimeter wave environments,seriously affecting the deployment of high-precision positioning services.Therefore,the traditional low-frequency wireless positioning technology cannot be directly applied to the millimeter wave frequency band,and how to use millimeter waves for high-precision wireless positioning has become a key research issue in the academic and industrial fields.Through research on wireless positioning technology in the mm-wave frequency band,this thesis summarizes and analyzes the challenges and key technologies from multiple perspectives.To address the problems of existing wireless positioning technologies in the millimeter wave frequency band,the following three key areas of research will be conducted.(1)In view of the fact that many interference nodes exist in the mmwave frequency band,which seriously affect the transmission and detection of positioning signals,this study proposes to configure different spatial relationships for multiple positioning signal resources in the positioning signal resource set.The receiver dynamically adjusts their spatial relations in real time according to the measurement results of the positioning signals,so as to avoid potential interference nodes as much as possible.At the same time,this paper sets different adjustment schemes according to different actual environmental conditions to achieve accurate uplink positioning in various channel conditions,which effectively reducing the impact of other transmission links on the current device positioning.The simulation verification results show that the proposed method can improve the measurement accuracy of the positioning signal by nearly 27%,significantly reducing the positioning error in the millimeter wave frequency band.(2)In view of the random variability of the channel conditions in the mm-wave band,only improving the transmission power can not significantly improve the reception quality of the positioning signal.This study proposes a mechanism for the joint positioning base station selection and power control for the terminals in the mm-wave heterogeneous network,and the proposed improved DQN algorithm is deployed in the intelligent macro base station,who can obtain the global environmental conditions in time to make centralized intelligent decisions.In addition,in order to reduce the signaling overhead and the implementation complexity of the uplink positioning mechanism,this paper proposes selective decision optimization under specific channel conditions.The simulation has verified that the intelligent macro station can converge to the optimal strategy after a certain number of training cycles,making the positioning energy efficiency of the millimeter wave terminal approximately 5 times that of the traditional UTDOA mechanism,and the terminal positioning power consumption is approximately 85%of the original mechanism.(3)This study proposes a multi-agent collaborative decision-making algorithm based on DDHQN to address the problem of multiple terminals choosing the same base station to assist in autonomous positioning in the millimeter wave frequency band,which results in significant fluctuations in terminal positioning performance and poor stability;Construct the decision-making system for collaborative positioning of multiple intelligent terminals as a competition plus cooperation model,and encourage intelligent terminals to participate in collaboration through the design of reward functions;In order to avoid decision conflicts between terminals,a record space representing decision trajectories is introduced to estimate the behavior of other terminals.The content of the record space is iteratively updated based on the wireless perception results of the intelligent terminal;Finally,the simulation proved that the proposed multiagent collaborative decision-making algorithm based on DDHQN enables each terminal to independently learn with the goal of maximizing group performance,and the maximum group reward that converges to is approximately 155%of the original DDQN algorithm.Through a comprehensive and profound analysis of wireless positioning technology in the millimeter wave frequency band,this study provides an intelligent solution for the deployment of high-precision positioning services in future mm-wave networks.The research results validate that the proposed spatial relationship configuration mechanism for positioning signals can significantly improve the positioning accuracy of the millimeter wave frequency band.The proposed joint positioning base station selection and transmission power control can significantly improve the energy efficiency of terminal positioning in the millimeter wave frequency band.The proposed distributed decision-making algorithm for multi-intelligent terminal collaborative positioning can effectively improve the positioning accuracy of the terminal group while reducing positioning power and signaling costs. |