Aiming at the problems of poor reliability and high false alarm rate of the fault detection models of bearings caused by the differences in working environment and equipment status, a multi-scale attention deep domain adaptation model was proposed according to the characteristics and needs of incipient fault detection. First, the monitoring signal was pre-processed into a three-channel data consisting of the original signal, Hilbert-Huang transform marginal spectrum and frequency spectrum. Second, the filters of different sizes were added into the residual attention module to extract multi-scale deep features, and the convolution-deconvolution operation was used to reconstruct the input information in order to obtain attention information, then a multi-scale residual attention module was constructed by combining the attention information and multi-scale features and was used to extract the attention features with stronger ability of representing incipient faults. Third, a loss function based on the cross entropy and Maximum Mean Discrepancy (MMD) regularization constraints was constructed to achieve the domain adaptation on the basis of the extracted attention features. Finally, a stochastic gradient descent algorithm was used to optimize the network parameters, and an end-to-end incipient fault detection model was established. Comparative experiments were conducted on the IEEE PHM-2012 Data Challenge dataset. Experimental results show that, compared with eight representative incipient fault detection and diagnosis methods as well as transfer learning algorithms, the proposed method can obtain the reduction of 62.7% and 61.3% in the average false alarm rate while keeping the alarm location not delayed, and effectively improves the robustness of incipient fault detection.

As Docker container lacks the ability of limiting network bandwidth resources, a bandwidth control mechanism was proposed for Docker container network based on Traffic Control (TC). Firstly, based on the real-time monitoring mechanism of CGroups file system, Virtual File System (VFS) of Linux kernel was used as a medium to pass the network control parameters set when Docker container was created to the Linux kernel controller TC. Then, the Intermediate Functional Block device (IFB) module was introduced to archive uplink and downlink bandwidth control, and the parameters (rate, ceil and prio) were used to achieve idle bandwidth sharing and container priority control. Finally, the specific network limitations were conducted by controlling the TC, and flexible network resource control between containers was realized. The experimental results show that the proposed mechanism can effectively limit the actual container bandwidth within 2% fluctuation range in the container exclusive bandwidth scenario, and can precisely limit the network bandwidth of the container with average 0.5% error range in the shared idle bandwidth scenario. Meanwhile, the mechanism can flexibly manage resources based on priorities. With the advantage of providing a more native interface for Docker and requiring no additional tools, this mechanism can provide a convenient and effective solution for fine-grained elastic network resource control on Docker-based cloud platform.

For the Orthogonal Frequency Division Multiplexing/Offset Quadrature Amplitude Modulation (OFDM/OQAM) system has no cyclic prefix, it is sensitive to time error and has a high requirement for frequency offset estimation in fast time-varying ground-air channel with large Doppler frequency offset, an AutoCorrelation Estimation (ACE) time frequency synchronization algorithm for OFDM/OQAM system in ground-air channel was proposed. In the algorithm, the symbol timing was used to achieve fast acquisition and timing with fewer auxiliary sequences. The frequency offset estimation was carried out by optimizing the autocorrelation sequence and performing two autocorrelation operations. The final frequency offset was obtained by weighting and averaging the estimated frequency offset of the two operations. The simulation results showed that symbol timing correlation peak of the ACE increased 3 compared with the Modified Linear Square (MLS) and Training Sequence 2 (TR2) algorithm. There was a 10dB SNR (Signal-to-Noise Ratio) gain when BER (Bit Error Rate) was 10 ^-2 at the en-route state of ground-air channel, and there was a 3dB SNR gain when the system BER was 10 ^-3 at the arrival state of ground-air channel. The simulation results show that the ACE algorithm further enhances time frequency synchronization accuracy and bit error performance.

Focusing on the issue that the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm may introduce some false targets in moving target detection with the presence of multipath reflections, a descending dimension method for designed matrix based on LASSO was proposed. Firstly, the multipath propagation increases the spatial diversity and provides different Doppler shifts over different paths. In addition, the application of broadband OFDM signal provides frequency diversity. The introduction of spatial diversity and frequency diversity to the system causes target space sparseness. Sparseness of multiple paths and environment knowledge were applied to estimate paths along the receiving target responses. Simulation results show that the improved LASSO algorithm based on the descending dimension method for designed matrix has better detection performance than the traditional algorithms such as Basis Pursuit (BP), Dantzig Selector (DS) and LASSO at the Signal-to-Noise Ratio (SNR) of -5 dB, and the target detection probability of the improved LASSO algorithm was 30% higher than that of LASSO at the false alarm rate of 0.1. The proposed algorithm can effectively filter the false targets and improve the radar target detection probability.

Concerning covering loopholes and uneven local deployment, a Directional and Heterogeneous Precision Self-deployment Algorithm (DHPSA) was proposed. Autonomous deployment process was divided into two stages. Firstly, a node moved to the destination path by choosing the optimal route in real-time under virtual forces of neighbor node and specified path. Then, through autonomous rotation and autonomous moving, the location of the node was finely tuned under joint virtual force of neighbor nodes and the accurate coverage of the target path was realized finally. The contrast experiments show that, compared with the VFPSA (Virtual Force-based Precision Self-deployment Algorithm), the coverage rate of the proposed algorithm is increased by about 4.4 percent, the overlapping rate is decreased by about 3.4 percent, moving distance is reduced by about 2.1 percent and deployment time is reduced by about 4.3 percent. The simulation experiment results show that the proposed deployment algorithm can effectively increase the coverage rate, decrease the overlap rate and reduce energy consumption.

It is a challenge for person identification to find an appropriate person feature representation method which can reduce intra-personal variations and enlarge inter-personal differences. A deep network for person identification based on joint identification-verification was proposed to solve this problem. First, the deep network model for identification was used to enlarge the inter-personal differences of different people while the verification model was used for reducing the intra-personal distance of the same person. Second, the discriminative feature vectors were extracted by sharing parameters and jointing deep networks of identification and verification. At last, the joint Bayesian algorithm was adopted to calculate the similarity of two persons, which improved the accuracy of pedestrian alignment. Experimental results prove that the proposed method has higher pedestrian recognition accuracy compared with some other state-of-art methods on VIPeR database; meanwhile, the joint identification-verification deep network has higher convergence speed and recognition accuracy than those of separated deep networks.

Aimed at the problem that existing RFID (Radio Frequency Identification) group proof protocols are inefficient and easily encounter many attacks like replay, tracking and so on, this paper proposed a new group proof protocol based on secret key-sharing tree. This protocol designed a new secret group-proofing key construction based on secret key sharing scheme. The group-proofing key was divided many times into many sub-keys to creat a key tree. This method increased the complexity of the construction of the secret key, increased the difficulty of that attackers attempt to recover the group key and increased the security of tag's group proof. The reader interacts with each tag only once to authenticate its validity and collect the group-proof information. This protocol enormously increases the proof efficiency. Compared to the existing protocols such as Yoking-Proofs, ECC-based and Tree-based, this protocol has better security and higher efficiency.

Due to individual differences in the ability, learning objectives and learning time of the learners,a learning resource recommendation method based on Particle Swarm Optimization (PSO) algorithm was proposed to provide the learner with a personalized digital curriculum. The knowledge structure chart was constructed applying concept map and knowledge structure. The learning objectives and the ability of the learner were analyzed based on item response theory. The PSO algorithm was adopted to select the appropriate e-learning materials from a mass of candidate materials and the adaptive e-course was composed and recommended. The learning time limit was considered while initializing particles, and some unnecessary particles were filtered out to improve the efficiency of the algorithm. When determining the optimal solution, the Sigmoid function was used to fix the particle update velocity to ensure it effective. The experimental results show that with the increase of the number of iterations, the difference between the recommended content and the learner goal is 0. The difference between the recommended curriculum and the learner ability is 0.6, and the overall difference is 0.25.It shows that the convergency and efficiency of the proposed method and the selected learning materials can meet the demands of different learners.

Although the normal cloud model is universally used, it faces some difficulties when describing some monotonic rise/fall conceptions. This model also has big subjective influence under multiple conditions and large computation consumption. To overcome these shortcomings, a new kind of exponential cloud model was provided along with a weight based cloud reasoning algorithm. By splitting the multi-condition generator to several single-condition generators, the algorithm firstly used Analytic Hierarchy Process (AHP) method to get weight of each property, and then used them to calculate weighted average of single-condition generator output to quantitfy value. The validation and effectiveness of this method is checked through a comparison between fuzzy reasoning and stimulation of torpedo avoid system.

The authentication efficiency of tags is always an important factor that restricts the extensive application of Radio Frequency IDentification (RFID). But there is not a good method to solve the problem till now. On the basis of tree-based RFID protocol, this article shared the secret of each path with many portions using the secret-sharing scheme. A new secret tree was created and a secret-sharing-based protocol was proposed while still keeping the searching efficiency. After being analyzed, the protocol is proved to be secure and efficient, and it also solves the key-updating problem which has slowed the study of RFID system for a long time.