Search

The 5G application scenarios such as remote control, high-definition video, unmanned mine car and unmanned aerial vehicle in intelligent mine put forward new requirements for the uplink rate of wireless network. However, the current 5G network uplink rate is insufficient, which leads to the limitation of intelligent mine business. And the existing 5G network uplink rate enhancement algorithm has limited range to improve the uplink rate. In order to solve the above problems, this paper proposes an uplink rate enhancement algorithm for 5G network in intelligent mine. By using supplementary upload technology, the spectrum resource aggregation is realized by overlaying Sub-3 GHz low band on the high band of C-Band, and the uplink rate of the network is improved by frequency domain resource allocation and time domain resource scheduling. In the middle-near point area, when the base station performs uplink data scheduling, the user equipment uses the 3.5 GHz frequency band to send uplink data in the uplink time slot of the C-Band frequency band, and uses the 1.8/2.1 GHz frequency band to send uplink data in the downlink time slot of the C-Band frequency band. In the far point area, the 3.5 GHz frequency band uplink is limited, and the user equipment only uses the 1.8/2.1 GHz frequency band to send uplink data. The test results show that the algorithm improves the mine near-point area, middle-point area and far-point area uplink rate by 17%, 41% and 213% respectively, and the average network uplink rate is significantly improved.

The large-scale MMC (matrix multiplication computation), MIC (matrix inversion computation) and MDC (matrix determinant computation) algorithm in secure outsourcing computing is widely researched and applied, while its existing problems are also increasingly highlighted. The problem of hiding the number of zero elements in the original matrix is one of them. However, there are few researches on it, and the existing researches can only protect the position of zero elements. There is no encryption method aiming at protecting the number of zero elements and at the same time achieving high efficiency, which is very unsafe in the cloud computing. To solve this problem, from the perspective of algorithm, the encryption method of permutation matrix is improved, and a new algorithm of secure outsourcing MMC, MIC and MDC is designed. The algorithm extends the algebraic structure to the finite domain. Firstly, the initial matrix is permuted randomly, then transformed by the unimodular matrix, and the encrypted matrix is sent to the cloud server. The cloud computes it and returns the results to the client. The client decrypts and verifies it. Through theoretical proof, the algorithm not only protects the number of zero elements of the original matrix, but also achieves the purposes of correctness, privacy and verifiability. Finally, experiments show that the encryption method based on unimodular matrix is efficient.

This work studies the effective multicast of data from base station to multiple devices in a small area, and proposes a fixed-relay based multicast scheme and a dynamic-relay based multicast scheme. Furthermore, aiming to minimize the total energy consumption of all devices, this work analyzes and obtains the optimum relay and its transmit power. Compared to the traditional multicast scheme, the proposed two schemes greatly reduce the total energy consumption in a multicast cluster.

Aiming at the copyright protection especially for small amount of vector data,a watermarking algorithm based on compressive sensing and multi-segment quantization modulation have been proposed in this paper.In the procedure of watermark generation,binary copyright image is divided into blocks and measured by compressive sensing method to reduce the watermark redundancy.Measured results are then modulated into four values to reduce the watermark length and ensure the watermark recovery.In watermark embedding and detection,the principle of multi-segment modulation on the angle is studied by establishing the mapping relationship between watermark bit and modulation segment index.Hence this watermarking principle enables each vertex to carry more watermark bits in a robust way.Experiments have been conducted to verify the effectiveness of the proposed watermarking algorithm.By using the proposed algorithm,a binary copyright image is able to be embedded into the small amount of vector data,which consists of only 100 vertexes,and can be extracted clearly.The experiments also reveal that the watermark capacity of the proposed algorithm achieves above 7 bit for each vertex,which is higher than the comparison algorithms.Regarding the algorithms robustness,the proposed watermarking algorithm shows its superior robustness against attacks including vertex insertion,vertex deletion,vertex sorting,translation,rotation and scaling.In summary,the contradiction between high watermark capacity and small data volume has been solved by the proposed watermarking algorithm.

There is an exterior orientation difference in each satellite video frame, and the corresponding points have different image locations in adjacent frames images which has geometric distortion. So the projection model, affine model and other classical image stabilization registration model cannot accurately describe the relationship between adjacent frames. This paper proposes a new satellite video image stabilization method with geometric distortion to solve the problem, based on the simulated satellite video, we verify the feasibility and accuracy of proposed satellite video stabilization method.