【作者】 杨久龙；

【导师】 方世良；

【作者基本信息】 东南大学 ， 信息与通信工程， 2017， 硕士

【摘要】 本文针对声学多普勒流速剖面仪(简称ADCP)多模式、多应用场景的需求,提出了组合式ADCP的概念,并在基于TMS320C6678的多核并行平台上设计并实现了具有高实时性与高可扩展性的组合式ADCP信号处理系统。具体工作如下:1.根据ADCP的基本原理与实现过程中的难点,分析总结了 ADCP信号与数据处理的规律。概括起来主要有两点:高实时性要求与瞬时计算量的矛盾;控制流程复杂、计算参数多变。所以,可并发、高瞬时计算能力的ADCP信号处理系统设计需求迫切。2.分析并总结了传统ADCP的共性和特性,提出了基于多种发射信号组合、多种测流方式组合以及多种处理算法组合的宽场景适应性的组合式ADCP。3.在对回波处理流程进行功能分解并模块化的基础上设计了基于流水线思想的回波处理流程,较串行处理方式成倍提高了回波信号处理流程的处理速度。提出了基于单波束测流分段并行计算和多波束测流分波束并行计算的组合式ADCP并行化测流计算方法。与串行计算方式相比成倍提高了处理速度。4.基于TMS320C6678多核处理器,设计并实现了组合式ADCP信号处理硬件系统。采用模块化多线程的软件设计方法,实现了多核多线程ADCP并行处理,并具有良好的可扩展性。系统完成后进行了实验室水池实验,验证了系统的可行性。更多还原

【Abstract】 According to the needs from the acoustic Doppler profiler(ADCP)of multi modes and multi application scenarios,this paper put forwards the concept of combined ADCP,designs and implements the signal processing system of the combined ADCP with high real-time and high scalability on the basis of TMS320C6678 multi-core parallel platform.Specific work is as follows:1.According to the basic principles of ADCP and the difficulties in the implementation process,the rules of ADCP signal and data processing are analyzed and summarized.There are two main points of generalization:the contradiction between the high real-time requirement and the instantaneous calculation,the complex control process and the variable calculation parameters.Therefore,the design requirement of ADCP signal processing system with concurrent and high instantaneous computing capability is urgent.2.Based on the research and synthesis of the commonness and characteristics of the traditional ADCP,a combined ADCP with wide range of scenes based on the various transmit signal combinations,the multiple flow measurement combinations,and a variety of processing algorithms is proposed.3.On the basis of decomposing and modularizing the echo processing flow,the echo processing flow based on the pipeline is designed,and the processing efficiency of echo processing flow is obviously improved.A parallel computation method of ADCP flow measuring algorithm,which is based on the piecewise parallel computation of single beam flow measurement and the splitting parallel computation of multi beam flow measurement is proposed.The efficiency of the echo signal processing algorithm increases exponentially4.A combined ADCP signal processing hardware system is designed and implemented based on TMS320C6678 multi-core processor.The modular multi thread software design method is used to realize multi-core parallel processing of ADCP with multi thread,and it has good scalability.After the completion of the system,the laboratory pool experiment was carried out to verify the feasibility of the system.更多还原