多元传声器阵列结构优化及其定位算法-控制工程专业论文.docxVIP

摘 要 构建适应性强、精确度高的定位模型及定位算法是声源定位系统要解决的核心问 题。然而，目前对机器人听觉感知的研究主要侧重于对语音的辨识和理解上，在定位 模型及算法方面由于缺少机理性的研究。在实际应用中只初步实现了对远场声源的定 向，缺少对高精度定距以及近场声源定位的处理方法。在声信号处理领域中对这一问 题的研究由于受模型复杂度和传声器阵列结构（尺寸和维数）的限制，一直没能从理 论上突破这一瓶颈，这大大降低了声源定位系统的适应性和实用性。 本文研究了一种多元传声阵列结构优化及其定位算法，重点分析了不确定条件下 多元传声器阵列的适应性、适用性和准确性；研究了不确定条件下的传声器阵列的定 位模型与定位稳定性，并且提出了可行的定位模型及计算方法；设计了传声器阵列的 结构，通过时延估计的定位方法，采用五元阵列交叉定位的方法解决了声源定位系统 的适应性和实用性的问题；仿真验证了交叉定位算法的有效性并总结出一些定位规 律；借助移动机器人实验平台，实际运用了优化阵列及其定位算法，实验测试表明该 方法中距离对声达时间差误差不敏感，论文研究的五元阵列结构及定位算法可实现较 高精度的声源目标定位。 关键字：机器人听觉 多元传声器阵列 交叉定位算法 I II II ABSTRACT It is the key problem for the sound source localization system to build high accuracy and strong adaptability positioning model and algorithm. But currently, the research of the robot auditory focuses primarily on the identification and understanding of voice, however, rarely mechanistic studies can be found in terms of sound localization model and algorithm. What is more, in practical use, there are only research findings on far-field sound localization, lacking approaches about distance determining and near-field sound localization. In the field of acoustic signal processing, because of the restriction of the model and the sensor array structure (size and dimension), there is still no theoretical breakthrough on this issue, which greatly reduces the flexibility and practicality of the sound source localization system. This thesis studies a multi-microphone array on structure optimization and localization algorithm. Adaptability, applicability and accuracy of the multi-microphone array in condition of uncertainty are mainly discussed. The localization model, as well as the stability of the multi-microphone array, in condition of uncertainty, is analyzed, and a viable localization model and algorithm are proposed. This thesis uses a method of cross localization, solving the problem of adaptability and practicality for sound source localization system. The effectiveness of cross location algorithm is verified by simulation, meanwhile, s