In recent years, extensive research has been conducted on radio resource allocation (RRA) in vehicular networks. Many studies have employed multi-agent Deep Reinforcement Learning (DRL) as an effective approach for making decentralized RRA decisions in highly dynamic and uncertain vehicular environments. However, a systematic evaluation and comparison of various multi-agent DRL algorithms in vehicular contexts remain lacking. In this paper, we address this gap by framing RRA problems in Cellular Vehicle-to-Everything (C-V2X) networks as a series of multi-agent interference games with ascending complexity as more realistic factors are introduced. We benchmark performance of classical multi-agent DRL algorithms in these environments. Our results offer insights into the relative significance of different Multi-Agent Reinforcement Learning (MARL) challenges in C-V2X RRA tasks, along with a comparative evaluation of multiple algorithms.

Collective Perception Messages (CPMs), which carry sensor data about surrounding objects and vehicle kinematics, are exchanged among Connected and Autonomous Vehicles (CAVs) to enhance situational awareness. This paper proposes a Value of Information (VoI)-driven CPM generation mechanism that enables each CAV to proactively determine when to transmit a CPM and what content to include, based on its assessed value to neighboring CAVs. The VoI of perceptual information is first defined from the perspective of the receiving CAV, by analyzing the spatiotemporal correlation between perceived objects and the receiver’s driving context. Guided by this, each CAV performs a three-stage process to generate CPMs in order to enhance information exchange efficiency, thereby improving Cooperative Perception (CP) performance. Simulation results demonstrate that the proposed mechanism effectively improves perception accuracy and reduces the communication burden on V2X networks.

Collaborative autonomous driving holds significant potential to improve the performances of Connected Autonomous Vehicles (CAVs). This paper presents a communication-aware hierarchical driving control mechanism designed to operate under non-ideal Vehicle-to-Vehicle (V2V) communication conditions. To address the impact of delayed and partial observations of CAV, a state augmentation method is first introduced to convert the resulting random-delay partially observable Markov decision process (RD-POMDP) into a standard Markov decision process (MDP), enabling the application of deep reinforcement learning (DRL) algorithms with theoretical convergence guarantees. Based on this formulation, a hierarchical DRL framework is developed, comprising an event-triggered upper-level for driving behavior adaptation at a coarse time scale and a periodic lower-level for motion control at a fine time scale. A modified Twin Delayed Deep Deterministic Policy Gradient with Prioritized Experience Replay (TD3-PER) algorithm is used to train the lower-level motion control policy, while an option-critic framework is employed to train the upper-level behavior policy, leveraging the pretrained low-level policies. Simulation results demonstrate the effectiveness of the proposed mechanism in collaborative driving scenarios with imperfect V2V communications.

To achieve accurate awareness of channel condition, the access point (AP) of a Wi-Fi network has to collect channel state information (CSI) from stations (STAs) periodically. However, existing CSI feedback mechanisms in Wi-Fi are situation agnostic, leading to substantial overhead due to the lack of adaptability and intelligence under dynamic and complex environments. To address this challenge, a concurrent dual channel information feedback mechanism with improve situation-awareness is proposed based on need-driven AP-STA coordination, aiming to maintain the accuracy of collected CSI while dramatically reducing the feedback overhead. By analyzing the latency tolerance of the channel information to be gathered, this concurrent dual feedback mechanism consists of both a delayed channel feature information (CFI) feedback by data frame and an immediate CSI feedback via control frame. In the delayed CFI feedback, a STA collaborates with AP and proactively determines when and what content of CFI to be fed back to the AP. Specifically, the CFI represents channel statistical channel features, which are crucial for the AP to learn the evolving channel conditions. Then, CFI is transmitted to the AP by piggybacking it in the uplink data payload at cost of a certain delay. On the other hand, STA can also utilize the existing CSI feedback mechanism for immediate CSI feedback. With the situation-aware CFI updates from both feedbacks, AP can effectively infer the downlink channel pattern and adapt the time-frequency resolution of CSI feedback to reduce the overhead. Accordingly, a deep cooperative multi-agent reinforcement learning algorithm is proposed to enable a closed-loop coordination between STA and AP for feedback. Simulation results confirm the effectiveness of our proposed mechanism.

With the development of Internet of Things (IoT) technology, the wearable devices have been widely used in different fields. However, few studies have focused on the application of wearable devices and IoT technologies on student movement detection systems for Physical Education (PE). This paper mainly designs an Message Queuing Telemetry Transport (MQTT)-based student condition monitoring system for student physical status monitoring, where the network transmission from the perception layer to the application layer in a multi-user scenario is considered. The proposed system consists of a data acquisition module, a communication module for message transmission, and a data analysis module for application. The (MQTT) protocol, as a low-overhead and low-bandwidth-consumption instant messaging protocol, is applied to enable the data to be published to the application layer clients that has subscribed to the corresponding topics in real-time. During the experiment, each perception layer client publishes different amounts of data to the corresponding topics of the broker, simulating multiple users sending data to the broker, and it sends the received data to the application layer client through the data flow function. The results show that the MQTT protocol has low latency in multi-user situations. Also, MQTT is able to provide real-time and reliable messaging services to the wearable devices with minimal data volume and limited bandwidth, which reveals the feasibility of its application in smart education.

The need for the campus safety arises with the increase of the number of colleges and campus facilities, posing significant challenges to traditional campus surveillance system. As a result, it is necessary to build a reliable, cross-platform and extensible surveillance system for campus monitoring. Therefore, this paper mainly proposes a web-based campus surveillance system based on ZLMediaKit, which is an open-source and high-performance streaming service framework. The system consists of a push/pull streaming server and a management front-end page. The streaming server supports numerous streaming media protocols and massive client connections, while the front-end page can access to back-end video streams and real-time display. Experimental results show that our system is stable and capable of handling multi-terminal streams with small consumption of resources.

Joint utilization of orthogonal radio resources from multiple domains such as spatial, time-frequency, and delay-doppler domains has become an important paradigm to support diverse QoS requirements (higher datarate, higher spectral efficiency, and low latency) in beyond 5G, and 6G. However, due to higher carrier frequency (mmWave) communication with closely packed massive MIMO antennas, and high-speed mobility in future wireless channels, severe non-orthogonal interferences are dynamically induced in multiple domains which dramatically deteriorate the communication datarate of current OFDM systems. In high speed mobility scenarios, orthogonal time frequency space (OTFS) modulation scheme achieves better communication performance than OFDM at higher modulation cost. Based on these observations, this paper is motivated to propose a novel, situation-aware, cost efficient, switched modulation in spatial, time-frequency, and delay-doppler domains termed hybrid multi-dimensional modulation (H-MDM) scheme that jointly optimizes the radio resource separation to minimize the non-orthogonality degree in each domain, and thus achieves maximized communication datarate under dynamically varying non-orthogonality conditions in those domains. Simulation results validate that the proposed H-MDM achieves maximized datarate compared to state-of-art MIMO-OFDM, and MIMO-OTFS systems under such randomly varying non-orthogonality conditions. Furthermore, we demonstrate that the proposed H-MDM scheme is highly advantageous for high speed mobility, and massive MIMO communication.

The increasingly diversified Quality-of-Service (QoS) requirements envisioned for future wireless networks call for more flexible and inclusive multiple access techniques in 6G for supporting emerging applications and communication scenarios. To achieve this, we propose a multi-dimensional multiple access (MDMA) protocol to meet individual User Equipment’s (UE’s) unique QoS demands while utilizing multi-dimensional radio resources cost-effectively. In detail, the proposed scheme consists of two novel aspects, i.e., selection of a tailored multiple access mode for each UE while considering the UE-specific radio resource utilization cost caused by non-orthogonal interference cancellation; and multi-dimensional radio resource allocation among coexisting UEs under dynamic network conditions. To reduce the UE-specific resource utilization cost, the base station (BS) organizes UEs with disparate multi-domain resource constraints as UE coalition by considering each UE’s specific resource availability, perceived quality, and utilization capability. Each UE within a coalition could utilize its preferred radio resources, which leads to low utilization cost while avoiding resource-sharing conflicts with remaining UEs. Furthermore, to meet UE-specific QoS requirements and varying resource conditions at the UE side, the multi-dimensional radio resource allocation among coexisting UEs is formulated as an optimization problem to maximize the summation of cost-aware utility functions of all UEs. A solution to solve this NP-hard problem with low complexity is developed using the successive convex approximation and the Lagrange dual decomposition methods. The effectiveness of our proposed scheme is validated by numerical simulation and performance comparison with state-of-the-art schemes. In particular, the simulation results demonstrate that our proposed scheme outperforms these benchmark schemes by large margins.

Applying of network slicing in vehicular networks becomes a promising paradigm to support emerging Vehicle-to-Vehicle (V2V) applications with diverse quality of service (QoS) requirements. However, achieving effective network slicing in dynamic vehicular communications still faces many challenges, particularly time-varying traffic of Vehicle-to-Vehicle (V2V) services and the fast-changing network topology. By leveraging the widely deployed LTE infrastructures, we propose a semi-decentralized network slicing framework in this paper based on the C-V2X Mode-4 standard to provide customized network slices for diverse V2V services. With only the long-term and partial information of vehicular networks, eNodeB (eNB) can infer the underlying network situation and then intelligently adjust the configuration for each slice to ensure the long-term QoS performance. Under the coordination of eNB, each vehicle can autonomously select radio resources for its V2V transmission in a decentralized manner. Specifically, the slicing control at the eNB is realized by a model-free deep reinforcement learning (DRL) algorithm, which is a convergence of Long Short Term Memory (LSTM) and actor-critic DRL. Compared to the existing DRL algorithms, the proposed DRL neither requires any prior knowledge nor assumes any statistical model of vehicular networks. Furthermore, simulation results show the effectiveness of our proposed intelligent network slicing scheme.

To meet the ever-increasing communication services with diverse requirements, situation-aware intelligent utilization of multi-dimensional communication resources is becoming essential. In this paper, considering a time-division-duplex downlink cellular scenario, a deep learning-based framework for multi-dimensional intelligent multiple access (MD-IMA) scheme is developed for beyond 5G and 6G wireless networks to meet the real-time and diverse quality of service (QoS) requirements by fully utilizing the available radio resources in heterogeneous domains. To achieve intelligent operation of MD-IMA, the proposed deep learning scheme is achieved based on the convergence of long short term memory (LSTM) and deep reinforcement learning (DRL). Specifically, an LSTM neural network is used to predict the long-term network dynamics and inference changes in QoS requirements of the MD-IMA. Meanwhile, a deterministic policy gradient (DDPG) algorithm, a model-free DRL technique, is adopted to optimize the multi-dimensional radio resource allocation in real-time by dynamically following the fluctuations of the network situation. With the aid of the DDPG algorithm, radio resource management for MD-IMA can be achieved efficiently with reduced processing latency as compared to the conventional model-based approaches. Furthermore, the effectiveness of our proposed deep learning framework for MD-IMA is validated through real-world cellular traffic data-sets. The experimental results demonstrate that the proposed scheme can outperform state-of-the-art algorithms.

Network slicing is a key paradigm in 5G and is expected to be inherited in future 6G networks for the concurrent provisioning of diverse quality of service (QoS). Unfortunately, effective slicing of Radio Access Networks (RAN) is still challenging due to time-varying network situations. This paper proposes a new intelligent RAN slicing strategy with two-layered control granularity, which aims at maximizing both the long-term QoS of services and spectrum efficiency (SE) of slices. The proposed method consists of an upper-level controller to ensure the QoS performance, which enforces loose control by performing adaptive slice configuration according to the long-term dynamics of service traffic. The lower-level controller is to improve SE of slices, by tightly scheduling radio resources to users at the small time-scale. To realize the proposed RAN slicing strategy, we propose a model-free deep reinforcement learning (DRL) framework, which is a hierarchical structure that collaboratively integrating the modified deep deterministic policy gradient (DDPG) and double deep-Q-network algorithm. Specifically, the lower-level control problem is a mixed-integer stochastic optimization problem with multiple constraints. This kind of problem is hard to be directly solved by the exiting DRL algorithms, since it involves searching for the solution in a vast set of mixed-integer action space, which will induce unbearable computational complexity. Thus, we propose a novel action space reducing approach, embedding the convex optimization tools into the DDPG algorithm, to speed up the lower-level control. Furthermore, simulation results confirm the effectiveness of our proposed intelligent RAN slicing scheme.

By leveraging direct device-to-device interaction, LTE vehicle-to-vehicle (V2V) communication becomes a promising solution to meet the stringent requirements of vehicular communication. In this paper, we propose jointly optimizing the radio resource, power allocation, and modulation/coding schemes of the V2V communications, in order to guarantee the latency and reliability requirements of vehicular user equipments (VUEs) while maximizing the information rate of cellular user equipment (CUE). To ensure the solvability of this optimization problem, the packet latency constraint is first transformed into a data rate constraint based on random network analysis by adopting the Poisson distribution model for the packet arrival process of each VUE. Then, utilizing the Lagrange dual decomposition and binary search, a resource management algorithm is proposed to find the optimal solution of joint optimization problem with reasonable complexity. Simulation results show that the proposed radio resource management scheme can reduce the interference from V2V communication to CUEs and ensure the latency and reliability requirements of V2V communication.

To improve the energy efficiency of multi‐cell massive multiple‐input–multiple‐output system while guaranteeing the information transmission quality, this study proposes a dual‐layer coordinated beamforming scheme. In the proposed scheme, the beamformer at each evolved node B (eNB) is divided into a cell‐layer beamformer and a user‐layer beamformer. The cell‐layer beamformer is used to mitigate inter‐cell interference (ICI) by exchanging long‐term channel state information (CSI) among eNBs. The user‐layer beamformer serves user according to the local real‐time CSI at each eNB. On the basis of the dual‐layer structure, the cell‐layer beamformers, the user‐layer beamformers, and power allocation are jointly optimised in order to minimise the total transmit power across all the eNBs subject to the signal‐to‐interference‐plus‐noise ratio requirements and single‐antenna power constraints. To make the original problem solvable, the ICI is replaced by its upper bound. Then, the problem is partitioned into two convex sub‐problems, and two iterative algorithms are proposed in order to find the sub‐optimal solution to the original optimisation problem. Simulation results show that the proposed scheme performs better than two reference schemes including the existing zero‐forcing scheme and coordinative multiple point schemes.

Downlink non-orthogonal multiple access (NOMA), where users are paired as user set and multiplexed in the power domain, is a promising technology for fifth generation (5G) communication system. This paper studies joint optimization of user pairing and power allocation to maximize generalized proportional fair metric subject to transmit power constraints. This problem can be divided into two parts: user pairing and power allocation. In order to reduce the computation burden, we present a pre-defined multi-user pairing criterion to exclude user sets, which are unsuitable for multiplexing. Furthermore, for a given user set, a low complexity multi-user power allocation scheme is proposed by exploiting the convexity of the optimization problem. Simulation results show that the proposed user pairing and power allocation scheme can significantly enhance the downlink system performance and reduce complexity compared to the existing schemes in NOMA systems.

Non-orthogonal multiple access (NOMA) is a promising technology for the fifth generation. However, the increasing number of multiplexed users brings two challenges. One is the computational complexity of existing grouping schemes increases rapidly. The other is the performance of multiplexed users deteriorates sharply. To deal with these challenges, we propose a novel multi-user grouping scheme with low-complexity to select several UEs for multiplexing. Moreover, a corresponding power allocation scheme is proposed to guarantee the performance of the multiplexed users. Simulation results show that the novel multi-user grouping scheme can reduce the computational complexity at the cost of user fairness in contrast to the full search method. Besides, the proposed power allocation scheme can improve the spectrum efficiency compared with the existing power allocation schemes in NOMA systems.