Video transmission, as one of the indispensable core services in the Internet of Vehicles, is confronted with numerous challenges such as the dynamics in the Internet of Vehicles environment, the limited resources, and the demand of vehicles for high-quality user Experience (Quality of Experience, QoE). In real-time video streaming scenarios, multicast optimization strategies, including convex optimization, game theory, stochastic optimization, etc., usually only group based on channel quality without considering task types and priorities, which may lead to the mixture of urgent tasks and ordinary tasks, affecting QoE. Therefore, we propose a task-aware three-stage collaborative optimization framework (TARA-IoV). The objective is to prioritize critical mission video services under limited bandwidth while maximizing the overall QoE. The framework first performs dynamic and adaptive vehicle grouping based on multi-dimensional features including task type, priority, geographical location, and channel state. Second, it conducts QoE-driven video quality layer selection leveraging Scalable Video Coding (SVC). Finally, it employs a deep reinforcement learning agent to dynamically allocate bandwidth with explicit task priority awareness under resource constraints. Evaluations on a real-world vehicle trajectory dataset demonstrate that TARA-IoV achieves improved QoE performance and more stable video delivery compared with existing schemes.

Digital twin (DT) migration technology is pivotal for ensuring seamless synchronization between physical entities and their virtual counterparts. However, in complex urban environments, high device mobility, dynamic network topology, and uncertain wireless channel conditions often lead to suboptimal migration decisions. Such inefficiencies exacerbate resource contention and diminish migration timeliness, thereby increasing DT state deviation and compromising service quality. To address these challenges, this paper proposes an AP group-enhanced digital twin edge network model, where multi-antenna access points (APs) are integrated as relay nodes to optimize channel quality and transmission reliability. Building on this, a Group-Collaborative Multi-Agent Proximal Policy Optimization (GC-MAPPO) migration strategy is presented. The strategy formulates the migration problem as a partially observable Markov decision process (POMDP). Specifically, the K-means++ clustering algorithm is first employed to construct optimal AP collaborative groups for mobile devices; subsequently, the MAPPO algorithm is utilized to derive optimal migration policies in dynamic environments. Experimental results demonstrate that, compared to existing baselines, the proposed GC-MAPPO scheme reduces the average synchronization delay by 6.69% to 37.05% and decreases DT state deviation by 29.24% to 82.91%.

3D object detection is a crucial task for autonomous vehicles to perceive traffic environments. Existing methods typically employ tightly coupled feature fusion strategies with fixed sensor combinations, but it fails to adequately capture the modality-specific characteristics, resulting in suboptimal object detection performance. To this end, we propose Chameleon, a novel multi-sensor decoupling system capable of selecting the appropriate sensor combination for object detection across diverse traffic conditions, which consists of two key components: (i) anuncertainty-aware contribution calculationcomponent leverages the uncertainty perceived by sensors to predict the parameters of traffic factors, followed by evaluating their importance; (ii) amutual information-enhanced sensor combination optimizationcomponent leverages mutual information calculations to enhance the mixture of experts and improve the reliability of sensor combination predictions, followed by achieving adaptive multi-sensor decoupling to ensure accurate 3D object detection while reducing inference latency. We implement and evaluate Chameleon using the nuScenes and nuScenes-C datasets. The experimental results show that Chameleon achieves average improvements of 1.10% in the mAP and 0.61% in the NDS compared to the state-of-the-art method across various traffic scenes.

Millimeter wave radar has recently emerged as a promising modality for enabling pervasive gesture recognition while protecting user privacy. However, personalized user behaviors, interference from unexpected actions, and long-term variability in user gestures significantly degrade the accuracy of gesture and user identity recognition, thereby compromising the quality of user experiences. To this end, we designmmGeswith four key modules: (i) afine-grained feature extractorextracts micro-level features from the time-series radar data to identify users' personalized behaviors; (ii) auser-specific feature classifierextracts coarse-grained features from a global perspective, followed by analyzing the micro-details of gesture features to recognize the user; (iii) avoting-based multi-user recognizerretrieves all pre-trained models from the user model database, followed by obtaining the probability indicators of each to return recognition results; (iv) alifelong learning modelretains previous knowledge while adjusting its feature selection capabilities using newly collected data to adapt to gesture changes. We implement and evaluatemmGesusing three self-collected real-world radar datasets, demonstrating its superior performance compared to other state-of-the-art gesture recognition methods.

Infrastructure-assisted autonomous driving has become a new paradigm that enables autonomous vehicles to fuse sensor data and improve driving safety, where a key enabling technology for achieving this vision is to real-time and accurate registering 3-D mmWave radar point clouds between the infrastructure and the vehicle. To this end, we proposeArtemis, a novel lightweight system capable of achieving real-time registration with decimeter-level localization.Artemisconsists of three components: 1) a modal association-based salient object extraction component leverages the complementary advantages of cameras and radars to extract semantics and areas of salient objects for radar point clouds; 2) a salient object shape construction component extracts the shape contour of salient objects based on their inherent geometries; and 3) a contour-guided 3-D point cloud registration component combines two key strategies, keypoint matching strategy and early exit strategy, to quickly select keypoints and transformation directions for achieving accurate registration in real-time. We implement and evaluateArtemiswith two multiview datasets collected in the CARLA platform and campus. The experiment results show thatArtemisachieves an average registration error of 0.33 m within 32.26 ms.

With the wide application of emerging technologies such as ultra-high definition video in Vehicular Edge Computing (VEC), the massive heterogeneous video data generated by vehicles have put forward higher requirements for real-time performance, energy efficiency and accuracy of processing. However, higher video analysis accuracy often leads to an increase in delay and energy consumption. How to balance the relationship between the three is an urgent problem to be solved. Meanwhile, the balanced or fixed bandwidth allocation mechanism adopted by most studies often ignores the characteristic differences of video tasks, resulting in inefficient resource allocation. At the same time, the security risks in the Internet of vehicles cannot be ignored. In order to deal with these challenges, this paper proposed a distributed task offloading framework combining Analytic Hierarchy Process (AHP) and Deep Deterministic Policy Gradient (DDPG). An adaptive bandwidth allocation mechanism based on the characteristics of video tasks is designed, and an improved blockchain consensus mechanism is introduced to ensure the optimal offloading decision in a trusted environment. Experimental results show that compared with the existing offloading schemes, the proposed algorithm reduces the task offloading delay by about 7.54%, reduces the energy consumption by about 6.37%, and improves the accuracy of video analysis by about 5.02% while ensuring security.

Multi-modal data-driven object detection typically depends on large-scale datasets. However, under adverse weather conditions, these datasets often exhibit a long-tail distribution, resulting in a substantial decline in detection performance and posing risks to system safety in applications such as autonomous driving. To this end, we propose Cerberus, a real-time and high-accuracy object detection system, which consists of two key components: (i) asimilarity-driven proposal extractioncomponent, which utilizes pre-trained detectors and convolutional neural networks to enhance low-quality features and suppress noise across image, LiDAR, and radar data, followed by aiming to extract and refine multi-modal features, thereby improving the quality of object proposals; (ii) anadaptive weight-based box enhancementcomponent, which integrates a frozen-weight regionbased convolutional network sub-network with a sensor fusion mechanism, followed by refining bounding boxes by normalizing confidence vectors and dynamically weighting multi-modal features, ensuring accurate detection under adverse weather conditions. Experimental results demonstrate that Cerberus surpasses existing baselines, yielding performance gains of 2.7% in the mean average precision and 2.0% in the nuScenes detection score under adverse weather conditions.

Federated learning (FL) has been extensively utilized in distributed learning scenarios for the Internet of Vehicles (IoV). However, two key challenges exist: 1) gradients are frequently transmitted between vehicles during FL training, which reduces the communication timeliness between the traffic participants and 2) the fixed gradient compression method cannot sufficiently adapt to the dynamic IoV network topology. Therefore, we design an adaptive vehicle clustering method constructed according to multiple attributes, such as computational resources, communication distance, and latency. Accordingly, we propose a dynamic gradient compression strategy that filters similar local training models between vehicles and uses the Wasserstein distance to compute a sparsity threshold. This threshold acts as a dynamic compression factor that compresses gradient model parameters, reducing redundant parameter transmission. Furthermore, we conduct experiments using two datasets to evaluate the proposed strategy’s effectiveness. The compression ratio improved by 132- and 178-fold compared to the baselines, and the aggregated accuracy increased by an average of 10.13%. Additionally, experiments incorporating communication noise revealed that the aggregation model of the signal noise ratio is -29 dB.

Query exchange in the Social Internet of Vehicles (SIoV) can protect users’ trajectory information. However, this method lacks an appropriate incentive mechanism, which leads to cooperative users refusing to participate in query exchange. In order to provide cooperative users with incentives to participate in query exchange, this paper proposes a smart contract-based query exchange (SC-QE) trajectory privacy protection method. By creating a many-to-many smart contract, the method encourages the cooperative users to bid to the requesting users. Subsequently, in order to select a Best Similarity Deviation User (BSDU) for the requesting user to perform query exchange, the users in the smart contract are modeled as a weighted bipartite graph, and the matching between the requesting users and BSDUs is realized by means of a weighted bipartite graph best matching algorithm. Following successful verification of the query exchange transaction in the smart contract, the base station distributes rewards to the BSDU and uploads the query exchange transaction to the consortium blockchain. Experimental results show that compared with the deviation-based query exchange (DQE) method, the proposed method reduces the user processing time by 12% while increasing the continuous anonymous success rate by 29%. Therefore, the proposed method can reduce the service query time and improve the level of trajectory privacy protection.

With the emergence of a large number of computing resource-intensive applications and a variety of content delivery services, data in Internet of Vehicles (IoV) is exploding. In order to improve the service performance of IoV, Vehicle Edge Computing (VEC) accelerates the response process of content requests and reduces the backhaul burden of the base station by caching content at the nodes of the edge network. However, the existing caching strategies are usually affected by high computing and communication overhead, and can not well capture the dynamic changes and content popularity of the vehicle network. In order to solve these problems, we design a novel Cooperative Caching scheme by using Mobility Prediction and Consistent Hash for Federated Learning (called CMCF), which integrates mobility prediction and consistent hashing into the content caching scheme, and uses the federated learning framework to optimize the cached content, then we use deep reinforcement learning algorithm to develop the optimal cooperative caching policy to reduce the average delay of content transmission. Extensive simulation results prove the superiority of our method. Compared with other advanced caching schemes, CMCF can increase the cache hit rate by 8.7% and reduce the average content delivery delay by 17.8%.

With the rapid advancement of in-vehicle communication technology, vehicular edge caching has garnered considerable attention as a pivotal technology to improve the efficiency of data transmission. However, existing studies often overlook the issues of increased average content access latency and decreased caching hit rate, stemming from the conflict between limited storage space in in-vehicle edge servers and vehicle mobility. To address these issues, this paper proposes a Multi-agent Federated Deep Reinforcement Learning based Collaborative Caching Strategy (MFDRL-CCS), leveraging Vehicle-to-Vehicle (V2V) communications. Specifically, we first perform vehicle connectivity prediction based on Recurrent Neural Network (RNN) considering the characteristics of vehicle nodes and their interrelations. Then, the optimal caching vehicle is selected based on the connectivity between vehicle nodes and the density of vehicle nodes. Meanwhile, a Multi-Head Attention Popularity Prediction (MHAPP) model is also constructed, which amalgamates multi-dimensional features, including historical popularity, social relationships, and geographic location, to predict content popularity. Finally, the edge collaborative caching model is formulated as a Markov Decision Process (MDP). Under the multi-agent competitive deep Q-learning framework, each vehicle learns the optimal caching strategy through an independent Q-network to maximize long-term rewards, and uses federated learning to train the caching replacement algorithm in a distributed manner. Compared to existing caching policies, the caching policy proposed in this paper improves the caching hit rate by approximately 19.8% and reduces the content access latency by about 12.5%.

User alignment refers to linking a user's accounts across multiple social networks, which is important for studying community discovery, recommendation systems, and other related fields. However, existing methods primarily perform user alignment by correlating user features, neglecting the causal relationship between network topology and user alignment, which makes it challenging to achieve superior user alignment accuracy and generalization capabilities. Therefore, we propose a counterfactual inference-based social network user-alignment algorithm (CINUA). This improves user connection retention due to the non-Euclidean geometric characterization of hyperbolic spaces. The similarity of aligned users is augmented using a hyperbolic graph attention network. User-feature embedding and fusion facilitate user relevance mining. Furthermore, there are causal relationships between network topology structure and user linkages. In various communities, there are some highly similar user pairs, and based on counterfactual inference, the network topology is adjusted to enhance sample diversity. Multilevel factual and counterfactual networks are constructed through iterative diffusion based on user alignment and their linkages. By integrating the users’ causal features in multiple networks, the accuracy and generalization capabilities of the user alignment model are effectively improved. In this article, the experimental results indicate that CINUA achieves a user alignment accuracy improvement of 5.98% and 3.03%, on two datasets respectively compared to the baseline methods on average. CINUA can achieve favorable alignment results even when the training dataset is small. This demonstrates that our algorithm can ensure both user alignment accuracy and generalization capability.

As the industry related to intelligent vehicles becomes increasingly mature, emerging in-vehicle applications and services are mushrooming. They have strict requirements on service response delay and network bandwidth. In the face of dynamic environment in the Internet of Vehicles (IoV), how to design an adaptive edge caching strategy becomes a challenge. To cope with this challenge, some researchers have introduced optimization methods based on learning algorithms in cooperative caching. However, general learning algorithms tend to waste bandwidth and computing resources on repetitive task. To this end, we make full use of the idle resources in the road to build a cooperative caching system and propose a Social-Aware Decentralized Cooperative caching (SADC) for IoV. This strategy uses the federated learning framework to train the collaborative caching algorithm based on Deep Reinforcement Learning (DRL). Among them, the Road Side Unit (RSU) is responsible for the training and updating of the global model, and vehicles use local data to provide local updates to the RSU, which then averages the updates provided by all vehicles to improve the shared model. In addition, we use the social network of vehicle users to obtain vehicle contact rates in different areas. The SADC strategy can reduce the content transmission latency and response time, thereby improving the experience quality of vehicle users. Compared with traditional caching strategies, this strategy reduces the average content access delay by about 20%. We also demonstrate the effectiveness of this strategy using an extensive set of experiments.

The application carrier of wireless body area network (WBAN) is human; due to changes in people’s sports status or physical health and other reasons, the business traffic fluctuates greatly, which requires the network to have good adaptability. In addition, the energy consumption problem is also a key factor restricting the applications of the WBAN. At present, the proposed MAC protocol is not highly adaptive and has low energy efficiency. To solve this problem, this paper proposes an adaptive MAC protocol based on IEEE802.15.6 for WBAN (A-MAC). The protocol sets the data to three priorities according to the type of service; the superframe structure of IEEE802.15.6 is improved and reorganized into four phases: the beacon phase, the contention access phase, the noncontention access phase, and the inactive phase. The length of the contention access phase and the noncontention access phase is adjusted according to the proportion of nodes that generate each priority data. The contention access phase is further divided into three subphases, and the length of the subphase is dynamically adjusted according to the data priority. In the contention access phase, all nodes compete for access channel according to the channel access policy. The random data that competes successfully transmits data directly, and the periodic data that competes successfully transmits data in the allocated time slots of the noncontention access phase. Finally through the simulation of the proposed A-MAC protocol and IEEE 802.15.6 MAC protocol and CA-MAC protocol in network performance which were compared, the results show that in terms of throughput, power consumption, and the network time delay, the network performance using A-mac protocol is better than the network performance using IEEE802.15.6 MAC and CA-MAC protocols.

In cognitive radio network, the activities of primary users will cause great influence on the stability of multiple hops routes between cognitive users. In this regard, a joint channel selection and routing protocol, termed as CSRP, is proposed to ensure route stability and reduce route latency between cognitive users. The channel availability based on historical information and the channel switching delay are used as the channel selection criteria to choose the end‐to‐end shortest route which possesses high data delivery probabilities and low delays. Besides, simulation results show that the proposed protocol has a better performance in terms of packet transmission delay and data delivery rate compared with the routing protocols based on delay (TDRP) and based on joint routing and channel allocation (PUB‐JRCA).