Keynote Speakers/主讲嘉宾

Keynote Speakers/主讲嘉宾

Oral Speakers


Junhui Fan - Information Engineering University

樊军辉 - 信息工程大学

Title: Underwater Acoustic OFDM Channel Estimation with Unknown Sparsity

Abstract: Underwater acoustic channels estimation accuracy seriously affects the demodulation performance of the receiver in underwater acoustic (UWA) orthogonal frequency division multiplexing (OFDM) communications. Many channel estimation algorithms based on compressed sensing (CS) have been proposed. However, these algorithms are usually sparsity demanding while the information of sparsity is unavailable in many UWA OFDM communications. In this paper, we propose a channel estimation algorithm based on fractional Fourier transform (FRFT) and orthogonal matching pursuit (OMP) technology, FRFT is introduced to process the synchronization signal so that sparsity is estimated. OMP is adopted to reconstruct channel impulse response (CIR), the most innovative feature of the proposed algorithm is the estimation of sparsity, which makes OMP algorithm more practical. Simulation results show that the proposed algorithm outperforms the least square (LS) and in UWA OFDM communication.


Zongfu Xie - Information Engineering University

谢宗甫 - 信息工程大学

Title:Intelligent Loading and Unloading System for Signal Processing Based on OBDD

Abstract:In order to realize the fast dynamic loading and intelligent switching for application tasks in the signal processing platform, this paper analyzes the system of intelligent loading and unloading, and builds the software and hardware resource model for the signal processing platform, situation parameter definition dictionary and intelligent loading and unloading rule base. Then through the intelligent rule decision algorithm based on the ordered binary decision diagram, the intelligent loading and unloading for the application task under the minimum manual intervention is realized. Finally, the algorithm proposed in this paper and sequential rule storage algorithm are compared. The results show that the algorithm of the intelligent loading and unloading for signal processing tasks designed in this paper has higher practicability, timeliness and intelligence.


Chuanqi Zhao - Information Engineering University

赵传奇 - 信息工程大学

Title:Design and implementation of virtual monitoring system for distributed heterogeneous signal processing platform

Abstract:There are many kinds of resources, complex structure and different standards in distributed heterogeneous signal processing platform system, so it is difficult to achieve unified management. Based on this, this paper designs and implements a high availability virtual monitoring system with resource networking and modelling. The system uses the hierarchical multi domain form to model the system network, and uses all levels of managers to manage them step by step, so as to realize the dynamic supervision of the whole system. The test results show that the proposed virtual supervision system can effectively and dynamically monitor the resources in the management system in real time, enhance the reliability and maintainability of the platform, and improve the resource management efficiency of the distributed heterogeneous signal processing platform.


Yibin Wu - Information Engineering University

吴艺彬 - 信息工程大学

Title:A Joint Implementation for Timing synchronization and matching filtering through the frequency domain

Abstract:In order to reduce the computational complexity of the high-speed satellite communication system, the article proposed an architecture that combined the timing synchronization and the matching filtering without mixing. In the architecture, the filter module and the synchronization module were improved to ensure that the filter operation results could be used in the timing synchronization module. In other words, the architecture could complete signal filtering and synchronization processing through a Fast Fourier Transform. So this method could not only reduce system complexity but also reduce the amount of computation. Theoretical analysis and simulation experiments show that the architecture has better estimation performance at low signal to noise ratio and high latency environment. At the same time, this architecture effectively reduces the amount of system calculations by reusing resources between modules. Compared with the Gardner algorithm, the scheme used in this paper can reduce about 70% of the calculation, which is suitable for high-speed demodulation system.

Chi Wei - Information Engineering University

魏驰 - 信息工程大学

Title:A Blind Separation Algorithm of PCMA Signals Based on MS-Gibbs Algorithm

Abstract:A blind separation method of PCMA signals with different symbol rates based on MS-Gibbs (multiple states Gibbs) algorithm is proposed. The prior probability of input symbol pair is calculated based on channel state and two input signal components, which is used to guide the update of symbol sequence. Simulation results show that the performance of the proposed algorithm is similar to that of DG-PSP algorithm, but the complexity of separation is greatly reduced through the proposed algorithm. Compared with no iteration in the 10−2 order of magnitude of BER (bit error rate),the proposed algorithm can obtain nearly 2dB SNR (signal-noise ratio) gain after 2 iterations. If the number of iterations is 4, the proposed algorithm can obtain nearly 4dB SNR (signal-noise ratio) gain.

Jianglin Yuan - Information Engineering University

袁江林 - 信息工程大学

Title:Recurrent Convolution Attention Model (RCAM) for Text Generation based on Title

Abstract:Natural Language Generation (NLG) is one of the most important part in Natural Language Processing (NLP). Recently, generating text automatically with deep learning method has been improved a lot. While there are lots of defects in text generation such as the quality is not satisfied and the text of title is not clear. The paper used the recurrent convolution attention model with LSTM (Long Short-Term Memory) cells for text generation by giving a title. The result proved that it can generate sentence according with the title and make the text express more fluently. Moreover, it uses less time to train by contrast with the SeqGAN (Sequence Generative Adversarial Networks). At the same time, the result is better than other attention mechanism with LSTM models. Therefore, it has more significance for NLP research.