Unmanned Aerial Vehicle (UAV) cooperative control means that a group of UAVs based on inter-aircraft communication complete a common mission with rational division of labor and cooperation by using swarm intelligence as the core. UAV swarm is a multi-agent system in which many UAVs with certain independence ability carry out various tasks based on local rules. Compared with a single UAV, UAV swarm has great advantages such as high efficiency, high flexibility and high reliability. In view of the latest developments of UAV cooperative control technology in recent years, firstly, the application prospect of multi-UAV technology was illustrated by giving examples from the perspectives of civil use and military use. Then, the differences and development statuses of the three mainstream cooperative control methods:consensus control, flocking control and formation control were compared and analyzed. Finally, some suggestions on delay, obstacle avoidance and endurance of cooperative control were given to provide some help for the research and development of UAV collaborative control in the future.

The particularity of Chinese data leads to noise information generation in the process of discrimination, and the traditional Convolutional Neural Network (CNN) cannot deeply mine emotional feature information. In order to solve the problems, a Dual Channel Gated Convolutional Neural Network model with Sentiment Lexicon (DC-GCNN-SL) was proposed. Firstly, the word sentiment score of sentiment lexicon was used to mark the words in the sentences, so that the prior knowledge of emotion was obtained by the network, and the noise information of the input sentence was effectively removed in the training process. Secondly, a gated mechanism based on GTRU (Gated Tanh-ReLU Unit) was proposed while capturing the deep sentiment features of the sentences, and the text convolution operation of the two input channels was used to fuse the two features, control the information transmission, and obtain more abundant hidden information effectively. Finally, the text sentiment polarity was output through the softmax function. The experiments were carried out on hotel review dataset, takeaway review dataset and commodity review dataset. Experimental results show that, compared with other models of text sentiment analysis, the proposed model has better accuracy, precision, recall and F1-score, and can effectively obtain the emotional features of sentences.

Considering the problem of out-of-stock and inventory overstock caused by dual-channel supply chain inventory system, three dynamic optimization models of three modes: single control, centralized control and cross-replenishment control of dual-channel inventory were established under the condition that both online and offline channels are facing stochastic demand. Firstly, based on the dynamic differential equation of inventory, guided by the control theory creatively, and by means of Taylor expansion and Laplace transformation, the feedback transfer function of dual-channel inventory system was obtained. Secondly, considering the periodic interactions, upstream and downstream interactions and inter-channel interactions in the process of cross-replenishment’s purchase-sale-stock, delay control, feedback control and Proportion-Integral-Derivative (PID) control were used to construct a complex interactive system with two inputs and two outputs, so as to explore the dynamic balance between supply and demand of the dual-channel inventory system itself and among channels, optimize the dual-channel inventory holdings, reduce the out-of-stock times and amount and keep them to a dynamic equilibrium. Finally, through numerical simulation experiments, three dual-channel inventory control strategies were compared. The simulation results show that when on online and offline channels were facing different distributions of stochastic demand, the residual stock of cross-replenishment control decreased by 4.9% compared with that of single control, and the out-of-stock rate of cross-replenishment control decreased by 66.7% and 60% respectively compared those of single control and centralized control. The experimental results show that when online and offline channels are facing different distributions of stochastic demand, the use of cross-replenishment strategy can effectively reduce inventory holdings, reduce the times and amount of out-of-stock, and thus save the inventory costs.

Focusing on the privacy security problem of Private Information Retrieval (PIR), a private information retrieval protocol based on point Function Secret Sharing (FSS) was proposed. The index of the retrieval was regarded as a special 0-1 point function, and the key group of the point function was generated by using the point function secret sharing technique, which was sent to p servers respectively. The retrieval results were obtained by XOR operation according to the responses returned by the p servers. The correctness, security and efficiency of the protocol were analyzed, which proves that the proposed protocol is secure and efficient. A concrete example was given to illustrate the validity of the protocol. Finally, the applications of the protocol to multi-term private information retrieval and keyword-based private information retrieval were introduced.

Aiming at the low efficiency problem of the existing granulation reduction algorithms for multi-granulation rough set, an Effective Algorithm for Granulation Reduction of Multi-granulation Rough Set (EAGRMRS) was proposed. Firstly, the lower approximation Boolean matrix of decision classes was defined by using the decision information system as the object. The defined matrix could be used for converting redundant and repeated set operations into Boolean operations in the process of granular reduction. Based on this matrix, the algorithm for computing lower approximation of decision classes and the algorithm for computing the important measure of granularity were presented. Then, focusing on the problem of redundancy calculation when computing the important measure of granularity, a fast algorithm for computing the important measure of granularity with dynamic increasing of granularity was presented. On the basis, the EAGRMRS was proposed. The time complexity of the proposed algorithm is O(| A|·| U| ²+| A| ²·| U|), in which| A|is the size of granulation set,| U|is the number of instances in decision information system. The experimental results on UCI datasets show that, the proposed algorithm is effective and efficient, the efficiency advantage of EAGRMRS is more obvious over Heuristic Approach to Granular Structure Selection (HAGSS) for multi-granulation rough set when the dataset increases.

Since the current network design of RS485 has limited the communication node number and scalability, a networking scheme for embedded control node cluster and a corresponding control protocol based on RS485 bus were proposed. By adding repeaters between the master and leaves, the networking scheme expanded the number of nodes on the RS485 bus to 27000, and constituted a cluster of embedded network nodes in one-to-many control mode. For this characteristic, the Modbus protocol was extended and the repeater layer stipulation was appended. Unlike traditional way that simply increases the length of the physical address, the new protocol used the method of local-addressing to break the restriction on the number of nodes which created by Modbus address length, and introduced the node scanning and error feedback mechanism. As a result, a reliable node control for the master controller was achieved. The control protocol not only keep the original simplicity of Modbus protocol, but also owns the features of portability and scalability, and it is easy to implement on a micro control unit. When the extended protocol was adopted, the turnaround delay increased by 10.36% with one layer of repeaters, while two layers of repeaters caused the delay increased by 69.9%. With two layers of repeaters, the total delay increased by 2.4 times than that of the original Modbus system. Nevertheless, the average delay in practical two-layer relay system was still controlled below 70 ms. The results show that the scheme can attain the goal of clustering management of embedded nodes at the expense of some real-time performance.

Concerning the low accuracy and convergence of Glowworm Swarm Optimization (GSO) algorithm when resolving high-dimensional functions, an Improved GSO (IGSO) algorithm with mutation operator and foraging behavior was proposed. Using mutation operator to guide the evolution of glow worms which cannot find their peers in the visible range, the proposed algorithm could enhance the utilization of outliers and improve the overall efficiency. The operator with foraging behavior substantially increased the accuracy and convergence speed by searching accurately in the global optimal field captured by the algorithm. In the meantime, the operator could effectively avoid local optimum and enlarge the global search range of the algorithm in the late stage. The experimental results indicate that IGSO has better ability of global optimization and higher success ratio than GSO according to the tests of eight Benchmarks.

A new method for matching feature points of images from two cameras of a stereo vision system in the game engine was presented. The main issue is how to meet the request of real-time processing and accuracy of the system. Our method is to calibrate the epipolar constrains and calculate the offset of two image along with the constraints of homography. Then we used a fast method to match the feature points in a local area. This method is proved to be real-time and accurate enough to fulfill the request of our game engine.

Multicast is the current prime network technique focusing on the group receivers,but the absence of security mechanism has limited its application in various network services.Group key management is one of the kernel problems in secure multicast.A group key management scheme for secure multicast was proposed,which can not only effectively reduce the key numbers of group controller,but also detract the probability of alliance attack.

An optimizing method based on immune genetic algorithm was presented for designing multilayer feedforward neural network. This algorithm could decide the structure of the multilayer feed-forward neural networks and search the proper weight of this network. The simulation experiments show that this algorithm has the better ability of convergent on whole solution space and the capacity of fast learning than genetic algorithm and momentum BP algorithm.

Based on ExLF Log file format, the Session Tabel was defined, and the conception of user session graph was expounded in detail. A heuristic algorithm of creating graph of user sessions identification was presented. At last, an example illustrated the efficiency of this algorithm.