Abstract High-precision classification of encrypted traffic plays an important role in ensuring the reliability and safety of intelligent connected vehicles. However, the communication environment of vehicles is affected by complex traffic scenarios and changing external environments, which introduces noise into the observed traffic (e.g., padding artifacts and retransmission bursts). In addition, there is a strong similarity between different service categories. Therefore, existing encrypted traffic classification techniques are not applicable. To overcome these challenges, we propose SpACNet, a collaborative CNN-Transformer dual-path spectrum sensing classification network. Specifically, in addition to using stream sequence information, SpACNet also uses layered multi-scale spectrum recalibration technology and gated axial self-attention mechanism for frequency-domain information to suppress the influence of aliasing artifacts and noise. In terms of feature fusion, orthogonal constrained dynamic tensors and gating mechanisms are used to integrate and balance time-domain, frequency-domain tensors, and interaction tensors. We evaluate SpACNet and three advanced baseline methods based on public and real-world datasets. The results show that SpACNet outperforms existing methods and demonstrates robust performance on datasets containing highly similar traffic categories. In addition, a series of ablation experiments is conducted to demonstrate the advanced nature of the proposed method.

For intelligent connected vehicles, dispersed computing can mine potential computing resources in the vehicular network for intelligent connected vehicles (ICVs) computation intensive and delay sensitive applications, and serve as a supplement and extension to share the computation workload when the edge is overloaded. To provide high-quality services, many existing studies show that dispersed computing has significant advantages in reducing latency and improving efficiency. However, existing work has clearly overlooked the dynamicity of vehicular networks and the latency impact caused by the heterogeneous computing capabilities of adjacent idle devices. In this article, we propose a new task offloading scheme based on dispersed computing to reduce overall task time. Specifically: 1) we model the minimization of total task execution time for multiple ICVs in a dispersed device-assisted network, comprehensively considering task computation time, transmission time, and waiting time; 2) an ICV scheduling algorithm based on swap matching is proposed to fully satisfy the quality of service requirements of ICV. Then, a Lagrange dual-based method is proposed to achieve computing resource allocation; 3) the experiments are conducted on real-world datasets. Compared to baseline heuristic schemes, our algorithm can effectively reduce the total execution time by an average of 4.8% to 31.8%. In special scenarios, the total execution time is reduced by an average of 10.7%. The proposed framework effectively mitigates performance bottlenecks caused by resource heterogeneity and network dynamics, providing a robust solution for vehicular computing systems.

Dispersed computing has provided an important computing paradigm for task offloading and resource allocation of intelligent connected vehicles (ICV), which can improve the processing efficiency of ICV applications. This paper investigates task offloading and resource allocation in ICV scenarios based on dispersed computing. To maximize the efficiency of the computation process of ICV applications, a two-stage optimization model based on a directed task graph is established. We have formulated the optimization problem of task offloading and resource allocation in dispersed computing networks. Then we prove that the proposed problem is NP-hard by reducing it to the soft capacity facility location problem. To reduce the complexity of solving, we decompose the proposed problem into two sub-problems: task offloading and resource optimization. We first propose a dynamic sorting (DS) algorithm to solve the task offloading problem, using the resources obtained from the previous iteration to solve the task offloading sub-problem and obtain the task placement result. Meanwhile, we consider two types of ICV applications with different processing efficiency requirements and design a low-complexity heuristic algorithm (Black-headed Gulls Optimization algorithm, BGO) to solve the resource optimization problem. Furthermore, we design the numerical simulations and scalability experiments to verify the effectiveness and scalability of the proposed scheme. Compared to baseline schemes, our algorithm can effectively reduce CPU utilization by 14.3%, memory utilization by 5.1% and processing time by 8.7%. In special scenarios, the processing efficiency is better than baseline schemes by at least 6.85% and 23.79%. The proposed scheme effectively mitigates performance bottlenecks caused by application heterogeneity and network dynamics, thereby improving resource utilization and processing efficiency for heterogeneous ICV applications.

In intelligent connected vehicle applications, tasks such as path planning and health management involve numerous matrix operations, particularly matrix multiplication. Due to limited resources, these tasks are often outsourced to the edge server. However, outsourcing these tasks involving matrix multiplication might incur potential risks, such as returning incorrect results to expedite processing or even exposing sensitive data during the computation. Privacy-preserving verifiable matrix multiplication schemes address these concerns. However, it is meaningful in practice only if the verification and decoding time is lower than that of local computation. In this paper, we propose a privacy-preserving verifiable matrix multiplication for intelligent connected vehicles that further reduces the verification and decoding time. To achieve this, we first reduce the length of the ciphertext of linearly homomorphic encryption when encrypting a group of messages. Subsequently, we construct our verifiable matrix multiplication scheme based on the improved linearly homomorphic encryption. It has a lower critical dimension than the state-of-the-art scheme with a similar security level, since the shorter ciphertext and the simpler linearly homomorphic encryption algorithm. Performance analysis and experimental results demonstrate that the critical dimensions of our improved scheme are reduced by 23.3%, while the communication cost is reduced by 68.3%, making it particularly suitable for intelligent connected vehicle applications.

Time series data, a crucial form of structured data across various domains, often necessitates a substantial quantity of labels for supervised learning. The complexity of annotating time series data, however, results in a high annotation cost. To address this challenge, active learning emerges as a strategic approach, selecting key samples and interacting with experts for data annotation. Its primary objective is to acquire high-quality training data at the lowest possible cost. This paper proposes a novel Semi-supervised ACTive learning (SACT) framework tailored for time series data. Initially, we establish a criterion to assess the importance of unlabeled samples, considering the unique characteristics of time series. We employ context embedding similarity scores to segment the time series data, prioritizing samples with high scores for labeling. Furthermore, to enhance the utilization of a single query, we propagate labels to adjacent data points. By experimenting on different combinations of query strategies and label propagation methods, our approach consistently outperforms other semi-supervised active learning frameworks for time series data in most scenarios. This superior performance underscores the efficacy of our approach.

The rapid development of the acrlong IoV has placed a heavy burden on the edge server. Conversely, many idle vehicles parked near the vehicle in demand are not utilized. Opportunistic computing of vehicles can organize these idle vehicles to provide computing services, but this also requires a secure data-sharing scheme to offer support. Although the existing ciphertext policy attribute-based encryption (CP-ABE) can provide a secure fine-grained data sharing, they either map the data to an algebra element that cannot be applied in practice due to the limited length, or they are hybrid schemes that cannot satisfy the forward security. In addition, they require a trusted authority (TA), which may be unrealistic in implementation. To cope with these issues, we propose a dual hybrid CP-ABE scheme without a TA for vehicular opportunistic computing (VOC) in this article. We exploit the dual hybrid mechanism to solve the problem of forward security in hybrid schemes and eliminate the TA by combining the characteristics of VOC. Then, we prove the acrlong IND-sCPA and forward security of the scheme. Finally, we evaluate the computation, storage, and communication cost from theoretical and simulation perspectives and compare them with other typical schemes. These results illustrate that the proposed scheme has better efficiency and lower storage and communication cost in general.

In wireless networks, introducing small cells into the macro cell can increase the system capacity, but may bring interferences into the original wireless networks. In this paper, we design an approach on the power control problems in the small cell networks, to control the interferences through control the transmission power. The interferences between the macro base station (MBS) and the small cell base stations (SBSs) are constructed as a Stackelberg game, where the MBS and SBSs act as the leader and followers respectively. Meanwhile, the status of the MBS and SBSs are described through two differential equations, to show the dynamic characteristics of the energy. Then the optimal control strategies for both the leader and followers can be given based on the open-loop Stackelberg equilibriums to the differential game. Numerical simulations and results show the effectiveness and advantages of the proposed algorithms.

Fog computing, as a new paradigm, has many characteristics that are different from cloud computing. Due to the resources being limited, fog nodes/MEC hosts are vulnerable to cyberattacks. Lightweight intrusion detection system (IDS) is a key technique to solve the problem. Because extreme learning machine (ELM) has the characteristics of fast training speed and good generalization ability, we present a new lightweight IDS called sample selected extreme learning machine (SS‐ELM). The reason why we propose “sample selected extreme learning machine” is that fog nodes/MEC hosts do not have the ability to store extremely large amounts of training data sets. Accordingly, they are stored, computed, and sampled by the cloud servers. Then, the selected sample is given to the fog nodes/MEC hosts for training. This design can bring down the training time and increase the detection accuracy. Experimental simulation verifies that SS‐ELM performs well in intrusion detection in terms of accuracy, training time, and the receiver operating characteristic (ROC) value.

In this study, the authors explore a joint cross‐layer optimised routing and dynamic power allocation for intermittently connected deep space backbone layer in deep space information networks via predictable contacts. First, the authors build up the payoff function and state dynamics based on differential game theory via two defined cost function paradigms. Then they establish a general differential game model for dynamic power allocation by the associated payoff function and state dynamics, and further propose the theoretical results of dynamic power allocation in a cooperative or non‐cooperative manner. To describe the routing metric, they introduce the concept of hybrid link homeostasis which illustrates the connection between the predictable contact and the transmitted power along the corresponding backbone link. In addition, they propose a polynomial time algorithm of cross‐layer optimised routing, which realises joint routing selection, transmitted power allocation and predictable contact schedule simultaneously. The numerical results demonstrate the effectiveness and feasibility of the authors proposed joint cross‐layer optimised routing and dynamic power allocation.

In this paper, a new construction of Z-complementary code sets is proposed based on conventional complete complementary codes, interleaving operations and orthogonality-preserving transformation. This method reduces the construction of Z-complementary codes to the construction of the shift sequences. The generated Z-complementary codes have a bivalued zero correlation zone (ZCZ). It is shown that, the number of mates of the new generated Z-complementary code sets can reach the bound P[N/Z] where P, N and Z denote the number of elementary codes, the code length and ZCZ length of the generated Z-complementary code sets respectively, and [x] denotes the largest integer no larger than x.