In order to improve the accuracy and speed of railway track detection, a new method of detecting left and right railway tracks based on deep Convolutional Neural Network (CNN) and clustering was proposed. Firstly, the labeled images in the dataset were processed, each origin labeled image was divided into many grids uniformly, and the railway track information in each grid region was represented by one pixel, so as to construct the reduced images of railway track labeled images. Secondly, based on the reduced labeled images, a new deep CNN for railway track detection was proposed. Finally, a clustering method was proposed to distinguish left and right railway tracks. The proposed left and right railway track detection method can reach accuracy of 96% and speed of 155 frame/s on images with size of 1000 pixel×1000 pixel. Experimental results demonstrate that the proposed method not only has high detection accuracy, but also has fast detection speed.

In the process of network traffic classification, the traditional model has poor classification on minority classes and cannot be updated frequently and timely. In order to solve the problems, a network Traffic Classification Model based on Ensemble Learning (ELTCM) was proposed. First, in order to reduce the impact of class imbalance problem, feature metrics biased towards minority classes were defined according to the class distribution information, and the weighted symmetric uncertainty and Approximate Markov Blanket (AMB) were used to reduce the dimensionality of network traffic features. Then, early concept drift detection was introduced to enhance the model's ability to cope with the changes in traffic features as the network changed. At the same time, incremental learning was used to improve the flexibility of model update training. Experimental results on real traffic datasets show that compared with the Internet Traffic Classification based on C4.5 Decision Tree (DTITC) and Classification Model for Concept Drift Detection based on ErrorRate (ERCDD), the proposed ELTCM has the average overall accuracy increased by 1.13% and 0.26% respectively, and the classification performance of minority classes all higher than those of the models. ELTCM has high generalization ability, and can effectively improve the classification performance of minority classes without sacrificing the overall classification accuracy.

Wrong object detection may lead to serious accidents, so high-precision object detection is very important in autonomous driving. An object detection method of Gaussian-YOLO v3 combining attention and feature intertwine module was proposed, in which several specific feature maps were mainly improved. First, the attention module was added to the feature map to learn the weight of each channel autonomously, enhancing the key features and suppressing the redundant features, so as to enhance the network ability to distinguish foreground object and background. Second, at the same time, different channels of the feature map were intertwined to obtain more representative features. Finally, the features obtained by the attention and feature intertwine modules were fused to form a new feature map. Experimental results show that the proposed method achieves mAP (mean Average Precision) of 20.81% and F ₁ score of 18.17% on BDD100K dataset, and has the false alarm rate decreased by 3.5 percentage points, reducing the false alarm rate effectively. It can be seen that the detection performance of the proposed method is better than those of YOLO v3 and Gaussian-YOLO v3.

TINY YOLO (TINY You Only Look Once) vehicle detection algorithm requires much amount of calculation which makes it difficult to achieve real-time detection in small embedded systems. Because plenty of zero values exist in a network weight matrix which makes the network a sparse structure, an improved version of TINY YOLO vehicle detection algorithm, called Xerantic-TINY YOLO (X-TINY YOLO), was proposed and accelerated in parallel way using architectural advantages of small Zynq SoC system. Original network structure of TINY YOLO was compressed and the operations of convolution steps were accelerated in parallel by using high efficient multistage pipeline. All multiply-add operations were concurrently executed within each stage of pipeline. By matching network structure, a method of data segmentation and transfer was also proposed. The experimental results show that, X-TINY YOLO only consumes 50% hardware resources on chip, and it can be implemented on small Zynq SoC systems which have higher performance-price ratio than GPU and CPU and is suitable for embedded implementation scenes. Its detection speed reaches 24 frames per second, which meets the requirement of real-time vehicle detection.

There are serious load imbalance and single point performance bottleneck effect in the traditional static topology leader-based distributed consensus algorithm, and the algorithm is unable to work properly when the number of breakdown nodes is larger than 50% of the cluster size. To solve the above problems, a distributed consensus algorithm (Yac) based on dynamic topology and limited voting was proposed. The algorithm dynamically generated the membership subset and Leader nodes to participate in the consensus voting, and varied with time, achieving statistical load balance. With removal of the strong constraints of all the majority of members to participate in voting, the algorithm had a higher degree of failure tolerance. The security constraints of the algorithm were reestablished by the log chain mechanism, and the correctness of the algorithm was proved. The experimental results show that the load concentration effect of single point in the improved algorithm is significantly lower than that of the mainstream static topology leader-based distributed consensus algorithm Zookeeper. The improved algorithm has better fault tolerance than Zookeeper in most cases and maintains the same as Zookeeper in the worst case. Under the same cluster size, the improved algorithm has higher throughput upper limit than Zookeeper.

Focusing on the problem that the privacy-preserving of identity authentication in Vehicular Ad Hoc NETworks (VANET), a conditional privacy-preserving authentication scheme was proposed. Firstly, this paper introduced the short signature technology, and then constructed a new identity-based short signature scheme. Compared with the well-known Conditional Privacy-Preserving Authentication Scheme (CPAS), the proposed scheme could reduce the computation costs required for both signature and verification processes and improve the communication efficiency. Secondly, the scheme divided the private signature key into two correlative sub-segments, so that it could effectively solve the issue of key escrow. Therefore, the scheme was especially suitable for the environment of VANET. Based on the proposed signature scheme, a conditional privacy-preserving authentication scheme was presented, which can achieve identity authentication with conditional privacy preservation. The theoretical and efficiency analysis shows that the scheme needs only three dot multiplication in the signature process and takes one dot multiplication, two pairing operation in the verification process. Especially, the proposed scheme use batch verification by adding the small coefficient test to accelerate the authentication speed and reduce the error rate.

In order to solve the safety problems of privacy in the processes of authentication and communication of Radio Frequency IDentification (RFID) system, a mutual authenticate and encrypt scheme with low resource consume, high-level security and applicable for most of RFID systems was designed. This scheme combined the improved Elliptic Curve Diffie-Hellman (ECDH) algorithm and Advanced Encryption Standard (AES) algorithm to implement functions of key distribution, certification and communication encryption. It used dynamic key to enhance security. In addition, this scheme reduced the operation scale with original security strength, and saved the overhead of system resources. The measured results show that this scheme can resist replaying attacks, impersonation attacks, man-in-the-middle attacks and Denial of Service (DoS) attacks so as to save system resources. It can be applied in the field of Internet of Things (IOT) which has requirements on security and costs.

Energy-efficient scheduling algorithms based on multiple Directed Acyclic Graph (DAG) fail to save energy efficiently, have a narrow application scope and cannot take performance optimization into account. In order to solve these problems, Multiple Relation Energy Optimizing (MREO) was proposed for multiple DAG workflows. MREO integrated independent tasks to reduce the number of processors used, on the basis of analyzing the characteristics of computation-intensive and communication-intensive tasks. Backtracking and branch-and-bound algorithm were employed to select the best integration path dynamically and reduce the complexity of the algorithm at the same time. The experimental results demonstrate that MREO can reduce the computation and communication energy cost efficiently and get a good energy saving effect on the premise of guaranteeing the performance of multiple DAG workflows.

Currently, most of proposed Radio Frequency Identification (RFID) authentication protocols cannot resist replay attack and altering attack. This article proposed a low-cost secure protocol, called Matrix-based Secure Protocol (MSP), which could resist these attacks. MSP utilized matrix-theory and Pseudo Random Number Generator (PRNG), and required only 1000 gate equivalents. Compared to previous proposed protocols using the same algorithm, MSP had less demand on the storage and the computing capability. Then, this article analyzed the security of MSP with Burrows-Abadi-Needham (BAN) logic. The conclusion is that MSP applies to RFID well.

The Camshift algorithm based on color-kernel can effectively track objects in a simple background, but it is easy to be interfered by illumination variation or the similar color object in the background. To improve the algorithms ability to respond to illumination variation, a multi-feature adaptive fusion scheme based on color, shape and texture was proposed. And further improvements have been proposed through modifying feature histogram and setting a reasonable search region to solve the problem of similar background. The experimental results show that the improved algorithm has higher tracking accuracy than traditional algorithm in the scene with illumination variation or similar background.

A method was proposed to incorporate semantic information based on TongYiCi CiLin and HowNet into tree kernel-based Chinese relation extraction, the impact of these two kinds of semantic information on Chinese entity relation extraction was compared and analyzed, and the interrelation between lexical semantic information and entity type information was explored. The experimental results show that this method can improve the performance of Chinese relation extraction in some degree, and TongYiCi CiLin can complement the entity type information to a certain extent. Therefore, no matter whether the entity type information is involved or not, its semantic information can significantly improve the extraction performance for most of the relation types, while some conflicts exist between HowNet and the entity type information, leading to its performance improvements only for several relation types when entity types are provided.