Collaborative perception by leveraging the shared semantic information plays a crucial role in overcoming individual limitations of isolated agents. However, existing collaborative methods tend to focus solely on spatial features information, while neglecting importance temporal dimension. Consequently, potential benefits collaboration remain underutilized. In this article, we propose Select2Col, novel framework that takes into account <underline...

Leveraging deep learning to detect malicious network traffic is a crucial technology in management and security. However, security has raised concerns among scholars. In this work, we explore executing targeted adversarial attacks for multi-classification detection with limited interactions. Specifically, constrain the number of interactions employ hop-skip-jump attack (HSJA) generate small samples. These samples are then heuristically used train generative (GAN) substantial quantity...

Cache-enabled small cell networks have been regarded as a promising approach for network operators to cope with the explosive data traffic growth in future 5G networks. However, user association and resource allocation mechanism has not thoroughly studied under given content placement situation. In this paper, we formulate joint optimization problem of mixed integer nonlinear programming (MINLP) aiming at deriving balance between total utility rates retrieved from caches. To solve problem,...

Privacy information theft traffic is usually detected using classification methods, and deep learning-based detection methods are effective for this task. However, these have complex preprocessing processes as well tend to ignore the features of network traffic, while generalization ability not outstanding. In paper, a model TCCN (Traffic Classification Capsule Network) based on capsule proposed. Meanwhile, PacketCGAN-based data balancing method introduced assist in classification. A new...

Dual connectivity (DC) is a promising solution to boost system capacity and enhance mobility robustness in the fifth generation networks, allowing mobile users (MUs) aggregate radio-frequency resources simultaneously from at least two kinds of serving base stations. In this paper, we focus on novel cache-enabled heterogeneous network, where MUs can flexibly schedule its traffic between traditional sub-6 GHz ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML"...

In the fifth-generation communication system, millimeter wave (mmWave) networks will coexist with traditional micro (μWave) networks, which allows for higher data transmission rate and better user experience. this paper, we give a comprehensive framework of mathematical models analytical methods multi-tier mmWave μWave hybrid caching based on stochastic geometry. We propose an association strategy by assuming average biased-received power Rayleigh fading. Accordingly, using D-ball...

Abstract Cache‐enabled networks are a promising solution to high video traffic, which allows popular files cache at the user equipment (UE) and base stations (BSs). To improve efficiency of file transmission, we adopt full‐duplex (FD) radio technology millimeter‐wave small BSs ( mm SBSs) in our network scenario. In this paper, users can get their required via one self‐communication, FD device‐to‐device (D2D) communication, SBS communication macro BS (MBS) communication. Different modes have...

Correntropy has been proved to be effective in eliminating the adverse effects of impulsive noises adaptive filtering. However, correntropy is not desirable when error between two random variables asymmetrically distributed around zero. To address this problem, asymmetric using an Gaussian function as kernel was proposed. always best choice and can further expanded. In paper, we propose a robust filtering based on more flexible definition correntropy, which called generalized that adopts...

Malicious traffic detection is important for the safe operation of cyberspace. Existing methods are difficult to extract discriminative features, leading rate bottleneck. In addition, performance significantly degraded in sample imbalanced scenarios, with poor generalization ability and insufficient scalability. Therefore, this paper proposes a multi-layer malicious model based on multi-branch octave convolution attention mechanism (MLMTD), which adopts multiple modules more diverse feature...

Outlier elimination is an important work for the trajectory data processing in vehicle test. The continuous outliers degrade accuracy of further and analyzing results especially. Kalman filtering least squares B-spline approximation are common method outlier elimination, but often cannot meet requirements. Robust M-estimation can obtain good its too complex. Hereby, a new based on Wright criterion was proposed to eliminate called odd-even extraction. extracted into 2 <sup...

In the fifth generation communication system, millimeter wave (mmWave) networks will coexist with traditional micro (μWave) networks, which allows for higher data transmission rate and better user experience. this paper, we give a comprehensive framework of mathematical models analytical methods multi-tier mmWave μWave hybrid caching based on stochastic geometry. We propose an association strategy by assuming average biased-received power Rayleigh fading. Accordingly, using D-ball...

The content retrieval delay is an important performance metric for enhancing user experience in wireless networks. In this paper, by modeling the locations of base stations (BSs) as Matern hard‐core point process type II (MHP), we analyze a typical cache‐enabled device networks under most popular policy. Since it intractable to get size distribution Voronoi cell MHP model, propose approximate formula based on empirical result Poisson and derive cellular load which denotes number devices...

Cache‐enabled heterogeneous cellular networks (HCNs) have been investigated extensively to alleviate backhaul congestion and reduce content delivery delay. In this paper, we jointly optimize placement user association minimize the average delay in cache‐enabled HCNs based on flow‐level models. This formulation considers (1) different timescales of delivery, (2) locality popularity, (3) heterogeneity spatial traffic distribution, which are often neglected existing researches. The joint...

An interpolation method based on bidirectional unequal interval kernel adaptive filtering was proposed to solve the missing data problem of trajectory measurement for vehicle test. The training samples filter are designed by before and after segment measurement. After training, is using predict data. To further improve precision prediction data, sample sampled at different intervals obtain time points. acquired, finely adjusted a higher result. Simulation processing results demonstrate...

Collaborative perception by leveraging the shared semantic information plays a crucial role in overcoming individual limitations of isolated agents. However, existing collaborative methods tend to focus solely on spatial features information, while neglecting importance temporal dimension. Consequently, potential benefits collaboration remain underutilized. In this article, we propose Select2Col, novel framework that takes into account \underline{s}patial-t\underline{e}mpora\underline{l}...

Traditional Encrypted Traffic Classification (ETC) methods face a significant challenge in classifying large volumes of encrypted traffic the open-world assumption, i.e., simultaneously known applications and detecting unknown applications. We propose novel Open-World Contrastive Pre-training (OWCP) framework for this. OWCP performs contrastive pre-training to obtain robust feature representation. Based on this, we determine spherical mapping space find marginal flows each class, which are...