With the rapid expansion of road traffic and vehicle scale, citizens have to face the serious life safety risks caused by vehicle accidents while enjoying an increasingly convenient life. Fortunately, Vehicular Ad Hoc Networks (VANETs) have provided us with important information about vehicle big data, which makes us have new methods to analyze vehicle traffic accidents. In this paper, we tackle the problem of predicting the risk of vehicle accidents and propose the trichotomy Adaboost with SMOTE and One-Hot encoding (AdaBoost-SO) algorithm to attain vehicle accident risk prediction model. In our work, predicting accident risk is mainly based on use of Big Data mining and analysis of real-life accidents data. Firstly, the experimental dataset is reconstructed by Synthetic Minority Oversampling Technique (SMOTE). We complement missing data and encode each sample feature to One-Hot code. Secondly, the trichotomy AdaBoost algorithm is respectively used to train a series of weak classifiers from the experimental dataset and then combine them into a strong classifier to get the prediction model. Finally, extensive simulation results illustrate that using the prediction model by trichotomy AdaBoost-SO algorithm can take the area under curve (AUC) of 0.77 and real-time into account. Through risk prediction and early warning, it provides the Intelligent Transportation System (ITS) and driving safety assistance with theoretical foundation.

In this paper, we study the edge caching problem by considering the heterogeneous context with unknown users’ preferences. The Cache Provider (CP) can personalize the users’ storage based on available data to maximize the overall cache hit rate, accounting for the dynamic natures of both mobile edge cache scenarios and users’ preferences. Toward this end, we introduce an online Bayesian clustering caching algorithm for the CP to autonomously learn the users’ interactive cache hit data in a collaborative way while maintaining sustainable scalability. Specifically, a Bayesian generative framework called the Dirichlet Multinomial Mixture (DMM) model is used to describe the uncertainty about the latent number of users’ clusters, each of which consists of the users with the same preference. Then, a dynamic clustering policy is proposed to obtain both the underlying mapping of users to clusters and the preferences of each cluster by using a collapsed Gibbs Sampling algorithm. Subsequently, cache decisions are made according to the generated mappings by extending the traditional cache bandit algorithm to a new bandit mechanism with clusters of arms, capable of expediting the learning process between exploitation and exploration. We theoretically characterize the value of dynamic Bayesian clustering for the long-term edge caching scenario with respect to the regret incurred by non-cluster schemes. Finally, using a real-world dataset, our numerical results show that the proposed scheme outperforms the caching algorithms without clustering in the uncertain network scenario.

Physical layer security and full duplex have been two widely applied techniques in wireless communications community, where the physical layer security can prevent eavesdropping without the upper layer data encryption, while the full duplex can improve spectrum efficiency through simultaneous transmission and reception. In this paper, we jointly consider full duplex and security by investigating the cross-layer optimization problem to maximize the secrecy rate in the multi-hop wireless network. This combination aims to improve both spectrum efficiency and security of conventional wireless systems at the same time. We first construct a cross-layer secrecy rate model based on full duplex constraints, secrecy capacity constraints, and secrecy flow balance. Then, we formulate it into a mixed integer and non-linear programming problem and reformulate it with reformulation-linearization technique and convex hull relaxation into the linear form. Finally, we validate our proposed optimization method and algorithm through comparing it with half duplex and jamming to demonstrate that the proposed combination of security and full duplex achieve the significant improvement of spectrum efficiency and security.

In this letter, we propose a pilot-assisted receiver scheme based on learnable successive interference cancellation (PA-LSIC) for uplink single-input multiple-output (SIMO) non-orthogonal multiple access (NOMA) systems. The PA-LSIC combines the successive interference cancellation (SIC) structure with the model-driven deep learning network. Considering the noise impact of channel estimation and the incomplete detection and cancellation in SIC process, we introduce some new parameters, such as noise cancellation factor and interference cancellation factor, which are optimized by using the back-propagation algorithm and random gradient descent algorithm. Numerical results show that the PA-LSIC has superior bit error rate (BER) performance and lower complexity during training and implementation.

Millimeter-wave (mmWave) bands are expected to be an important choice for future vehicular communication to support Gbps links for reliable data transfer in high-rate applications. The recent online learning technologies addressed the problem of fast beam tracking by exploiting user location information and mining received data in mmWave vehicular systems to adapt to the vehicle’s environmental situation. However, the fairness and efficiency over mmWave beams are difficult to maintain on the move, especially for high-density traffic, since the number of available beams is quite limited by hardware and cost for current antenna arrays. Fortunately, the social structure of preferences between the neighboring smart cars and their passengers can be leveraged to improve the beam coverage efficiency by performing the broadcast transmission via a single beam. In this paper, we propose a double-layer online learning algorithm, namely, context-and social-aware machine learning (CSML), that is based on the context and social preference information of vehicles and passengers, to realize fast beam access with broadcast coverage in mmWave communication systems. Based on the multi-armed bandit model, CSML embodies the selection of appropriate beams in the first layer and steers the broadcast angle along these beams in the second layer by aggregating the received data. Furthermore, CSML needs to adjust the timing of exploration and exploitation based on the social information, i.e., the probability of vehicles meeting with each other that have the same preference. Finally, we perform an extensive evaluation using realistic traffic patterns and show that CSML increases the efficiency of mmWave base stations by using social data and can achieve near-optimal system performance.

Accurate and stable dynamic target positioning can provide high-precision vehicle motion state information for the autonomous driving decision-making system, which plays a very important role in the development of autonomous driving technology. In order to further improve the accuracy and stability of dynamic target positioning, based on the combination of Global Positioning System (GPS) and Ultra Wide Band (UWB) ranging information, the Kalman filter (KF) algorithm is used to process the measurement data of dynamic target. Since it is difficult to determine the system model noise and measurement noise under dynamic conditions, this paper introduces adaptive factor and dynamic window adjustment based on the covariance matching method, then proposes an adaptive Kalman filter (AKF) combination positioning method integrating UWB and GPS. In this method, the GPS single point positioning based on the sliding window AKF algorithm of covariance matching is used when the target is in the linear motion state. Under the condition of nonlinear motion of the target, the UWB/GPS integrated location is on account of the adaptive unscented Kalman filter (AUKF) algorithm which imports the filter divergence criterion. The simulation results confirm that the combined positioning method can adapt to the dynamic positioning requirements in complex environment and outperforms the traditional KF. The developed filtering algorithm has high adaptability, strong robustness, and can effectively improve the positioning accuracy