声定位系统及其数字信号处理技术的研究

【作者】 曹国强；

【导师】 程翔； 杨亦春；

【作者基本信息】 南京理工大学 ， 机械电子工程， 2002， 硕士

【摘要】 智能雷弹是防御超低空飞行的武装直升机的重要武器系统，而被动声探测技术是其关键技术，本文结合“九·五”国防预研项目“反直升机智能雷弹声复合引信技术”的研究，对空间目标声定位算法和风对被动声定位的影响及如何对这种影响进行修正进行了深入分析，并在此基础上设计了DSP软件系统。 本文介绍了被动声定位的原理及空间声定位的算法，推导出了被动声定位公式，并利用声程差组合改进了被动声定位公式，提高了定位的精度。重点讨论了风对于被动声定位所造成的影响，并在此基础上推导出了修正后的定位公式，并进行了计算机仿真计算，对修正前后进行了对比分析，得到了比较满意的结果。 为满足被动声定位技术的实时性和精确性的要求，本系统采用了TI公司的TMS320C32浮点DSP芯片作为计算用CPU，对DSP的存储区进行了扩展，完成了被动声探测系统的DSP软件设计。该软件运行稳定，且满足精度和时间要求。 运用模式识别的理论，研究对目标特征提取和识别的方法。探讨了利用神经网络技术进行目标识别的方法。更多还原

【Abstract】 The intelligent mine will be the most important weapon system to defend arming helicopter hedgehopping in future war. Passive acoustic detection technology is the key technology. With the researching of the national defence pre-researching project which is the research of acoustics compound fuse of anti-helicopter intelligent mine, this paper presents acoustic localization algorithm and discusses the affect of wind to the localization, then a method of how to correct the affect of wind has been studied, upon this a DSP software system has been designed.The paper introduced the theory of passive acoustic localization and the algorithm of space acoustic localization. By use of combined acoustic path difference, the paper amended the passive acoustic localization formula and improved the accuracy of localization. The affect of wind to the localization was discussed specially, upon which a new formula was deduced, and we have tested this new formula by use of computer simulation, which has given a satisfied result.In order to satisfy the need of real time and accuracy of the passive acoustic localization technology, we use TMS320C32 (DSP) which produced by TI company as computing CPU, RAM & ROM of DSP had been extended largely. A whole software system of the passive acoustics detection system has been worked out. The software can satisfy the need of time and accuracy simultaneously, and worked stably.By using pattern recognition theory, we studied target characteristics extract and recognition, as well as neural net recognition algorithm.更多还原