This paper proposes a novel Particle Swarm Optimization-Aided Reinforcement Learning (PSO-RL) framework for efficient beam and power management in reconfigurable intelligent surface (RIS)-assisted millimeter-wave (mmWave) networks. To optimize dynamic beam coordination and resource allocation, we introduce a reinforcement learning (RL)-based approach that integrates value and policy networks, with the policy network enhanced by a target policy network (TPN). By employing PSO, we refine TPN parameters to accelerate convergence and enhance solution exploration. Simulation results validate the effectiveness of the proposed PSO-RL framework, demonstrating significant performance improvements in RISassisted network deployments. The method enhances system capacity by 29.75 % and improves edge user capacity by 36.65 %. These results highlight the potential of PSO-RL for optimizing beam allocation in dynamic urban network environments.

Reconfigurable Intelligent Surfaces (RIS) are perceived as cost-effective entities capable of expanding network coverage and enhancing system performance in 6G networks. Considering an RIS-assisted 6G millimeter network, this paper aims to find low-complexity beam and power allocations that maximize the network throughput. To resolve the complexity of such a mixed-integer non-convex programming problem, an iterative power decremental method is developed to solve the beam and power allocation problem alternatively under the system design constraints, making use of the concept of difference of convex functions and the zeroth-order gradient ascent method. Simulation results verify the superiority of the proposed method and its effectiveness in RIS-assisted millimeter networks.

Mobile networks have seen more than a thousand times traffic increases in the past decade. Network operators face an ever increasing demand and cost to deploy denser networks in order to maintain the quality of services. In view of this paradigm shift, an open cloud radio access network (C-RAN) service model is studied in this work. Integrating the cross-layer functions of spectrum resource sharing, channel and power allocation, and interference management in a C-RAN architecture, we explore the feasibility of providing a scalable yet efficient broadband wireless service with dense small cell networks (DSN). The proposed methods and architecture can be applied to DSN that support the functions of (further enhanced) inter-cell interference cancelation (feICIC/ICIC) techniques of 3GPP standard, and have the potential to provide a cost-effective solution for high-quality DSN. Simulation results across an area of 100 km$^{2}$show that the proposed scheme can offer an aggregate downlink throughput of 21 Gbps over a maximum channel bandwidth of 20 MHz, which is 5 times the throughput with a typical ICIC method. Moreover, the service satisfaction degree can reach 85% under the proposed C-RAN architecture, making it a promising and cost-effective solution for future broadband wireless services.

A streaming-based Indoor Positioning System (IPS) is developed for beam-based cellular networks that apply directional beam patterns for communications and sensing. The proposed IPS is designed to provide a real-time positioning service by leveraging the computing framework and the streaming modules of Apache Spark streaming. Under the Spark Streaming framework, a beam and model-based location tracking method is designed and optimized for positioning precision and resource utilization. Experiments show that the average positioning error of the proposed user tracking method is about 0.4 meters, which achieves 56.2% improvements over static locating methods. More importantly, a static timing analysis is carried out to evaluate the throughput and bottlenecks of the proposed IPS system. Based on the analysis, our proposed architecture is capable and promising for low-latency IPS provisioning under the O-RAN framework.

A framework that incorporates channel modeling, position estimation, and error analysis methods is developed for network-based indoor positioning with radio signal strength (RSS) measurements of WiFi access points (APs). To construct an accurate channel model with RSS measurements severely influenced by propagation attenuations, multipath reflections, and shadowing effects, a novel sparse Bayesian learning algorithm is developed to model the radio power map (RPM) in indoor space. Based on the proposed RPM model, a 2-stage positioning method is further developed. In the first stage for coarse positioning, the location is determined up to a room-scale indoor space. Then, in the second stage for fine positioning, the RPMs of the given indoor space are used for location estimation in the space with a Bayesian estimator. The mean squared positioning errors are verified with the Bayesian Cramer-Rao lower bound. Extensive experiments show that the average positioning error of the proposed RPM-based approach is 1.98 meters which achieve 22 percent improvements over the state-of-the-art RSS-based indoor positioning methods. More importantly, the proposed modeling and positioning method can effectively exploit the spatial relationship in the RSS samples to improve positioning accuracy.

A new approach for indoor positioning is presented, aimed at designing a WiFi positioning system that is feasible and convenient for both service providers and end users. In the proposed approach, only access points (APs) need to collect the received signal strengthes (RSS) of mobile devices, and use these RSS samples to jointly estimate the devices' locations. To enhance the accuracy of positioning, the relationship between the RSS samples and their geometrical locations is explored, leading to a sparse Bayesian model for the radio power map of the RSS observations of each AP. With more than 20 training anchors, the accuracy of the proposed model-based positioning method can be lower than 3.4 meters in an indoor space with only 4 randomly deployed APs, which outperforms the fingerprinting method by 0.4 meter. Extensive experimental results also verify that the proposed positioning service can offer considerable accuracy with only limited efforts in training, suggesting that the prototype is realistic for randomly deployed dense WiFi networks.