Modern airborne radar systems have been equipped with multi-dimensional and full-airspace capabilities for de-tection and tracking.With the increasing maneuverability of airborne targets and the complexity of the electromagnetic envi-ronment,target tracking,as a key radar technology,is facing ever-growing demands for tracking accuracy and real-time re-sponse capability.In response to these challenges,filtering algorithms for air-to-air multi-target tracking in airborne radar systems are investigated in this thesis.To address the low tracking accuracy of traditional Interactive Multiple Model(IMM)algorithms for maneuvering targets,an Adaptive IMM Particle Filter is proposed by correcting the Markov transition proba-bility matrix.Model matching and switching responsiveness are enhanced by the proposed method.Simulation results dem-onstrate that position tracking accuracy is improved by 16.1％,and velocity tracking accuracy is improved by 26.8％,com-pared to conventional approaches,highlighting its significant advantage in maneuverability adaptation.
Aiming at the fuzzy problem of the image restored by the electromagnetic leakage signal of the display, this article proposes a restoration method of electromagnetic leakage information based on the fusion of wavelet analysis and interframe interleaved image fusion algorithm. Based on the principle of multi-frame average denoizing method, it makes use of the characteristics that the display depends on the periodic repetition of different video frames when displaying image information. Successive frames are superimposed to achieve image enhancement, which effectively reduces the influence of environmental noise. At the same time, the fusion of one-dimensional wavelet transform technology further improves the quality of image reconstruction. The experimental results show that, compared with the direct image reconstruction, the image signal-to-noise ratio, contrast ratio and acuteness of the fusion algorithm using wavelet analysis and frequency domain hierarchical image reconstruction are improved by 79.7%, 15.9%, and 47.8%, respectively. The algorithm has significant application value in the field of electromagnetic leakage signal recovery and signal processing.
Due to the increasing risk of privacy leakage,the collected data usually contains a large amount of privacy information,data collectors are reluctant to share their private data,which leads to result in"data silos".Federated learning enables data sharing without leaving the local area,but there are still some problems on sharing among multiple data.On the one hand,the centralized processing of central server suffers from expensive cost and single point of failure.On the other hand,for multi-institutional data sha-ring,model training might be affected by participating nodes mixed with malicious nodes,which leads to data privacy leakage.There-fore,Based on above analysis,a new distributed federated learning architecture is proposed to combine blockchain and federated learn-ing,realize the efficient node selection and communication,and it enables direct communication between participation nodes instead of relying on central server.Based on the proposed architecture,a reputation-based node selection algorithm scheme(RBLNS)is pro-posed to screen the participating nodes,and ensure the privacy and security of the participating nodes.The experimental results show that the RBLNS significantly improves the test performance of the model.
Mongolian script conversion plays a pivotal role in promoting the application and dissemination of the Mongolian script,enriching the diversity of Chinese culture,and fostering economic prosperity in Mongolian regions.Aimed at the poor efficiency and accuracy of Mongolian script conversion,this paper proposes a method based on lightweight convolutional neural network(CNN)and FPGA hardware accelerator The CNN is used to carry out the handwritten Mongolian script recognition,and combined with the font conversion library,a simple and efficient Mongolian font conversion is achieved through the mapping relationship between the recogni-tion results and the fonts.This approach has the advantages over other methods in the efficiency of the CNN and FPGA hardware ac-celerator,improving the conversion efficiency while meeting the requirements of device cost,power consumption,and portability.The circuit of network model is designed and optimized by using Xilinx's XC7Z020CLG400-2 chip.On this basis,a handwritten Mon-golian font conversion system was implemented.Test results show an accuracy of 95.62％for converting the handwritten Mongolian script to the target font,with a conversion time of 1.43 ms,a power consumption of 0.341 W,and a peak throughput of 6.64 Gops for the accelera-tor.It is of significance for the research results to promote the preservation of Mongolian culture and economic development.
针对确定性机载网络通信系统实时性强、可靠性高的需求,设计了一种时间触发以太网终端的实现方案;以MicroB-laze软核处理器为核心,使用LWIP协议栈实现UDP通信;经过60s的性能测试,平均带宽约为621Mbits/s;在TTE交换机与终端设备之间,使用主从式同步方式实现时间同步;经过10次时间同步测试,链路传输延时在150～250 μs之间;建立时间同步后加入时间触发任务,进行周期性的发送时间触发消息;为了保证系统可靠性,在传输层中采用MD5数据校验,并设计了一种四余度管理机制;测试结果满足TTE终端的功能需求.
