室外安防小型智能声测传感节点测向方案研究

发布时间：2018-06-07 08:32

本文选题：智能传感器网络 + 智能传感节点　；参考：《南京理工大学》2017年硕士论文

【摘要】：随着社会的发展和科学技术的进步,数字网络生活时代已经到来,家庭智能化、物业管理现代化和社区服务信息化已成为发展趋势,智能安防的出现成为必然。本文针对室外远场环境,采用小型传声器阵列构成的测向系统作为智能传感节点,为智能传感器网络实现对声源位置的定位打下基础。智能声传感器网络采用分布式结构,具有成本低廉、部署简便灵活、精确度高等优越性能,可用于室外监控等方面。本文研究的测向系统包含三个部分,分别为传感、处理、通信部分,可实现对声音信号的采集、分离和声源位置测向等功能。其制成的智能传感节点外围有金属材料做成的封装外壳保护内部电路,可以在恶劣环境、天气下代替人工对室外环境进行长时间的监控。在智能声传感器网络结构中,考虑到环境监测无需实时处理,本文对所监测的声音信号采用断续实时处理模式。受系统结构小型化、处理器的运算性能等方面的限制,节点的数据处理量不能太大,需要在时间限制内完成计算任务。采用声音活动性检测(VAD)与广义互相关(GCC)相结合的方式,在节点处理器上对采集到的数据进行声音活动性检测截取出有效信号段,然后利用算法PHAT-GCC进行测向估计,后台可综合多个节点的测向结果最终实现联合定位。这种利用单节点测向、多节点定位的方式,减少了在节点上的处理量及在传输上的通信带宽。在单节点的测向过程中,节点由于受到结构小型化及其封装外壳的影响,在远场环境下得到的实际声源测向结果存在一定的偏差。文章基于上述问题做了多种实验并对结果进行分析,从中找出对误差的补偿方法,提高小型智能传感节点的测向准确度,为智能传感器网络在声源定位的实用化进程方向上做了铺垫。
[Abstract]:With the development of society and the progress of science and technology, the era of digital network life has arrived, the family intelligence, the property management modernization and the community service information have become the development trend, the appearance of intelligent security has become inevitable. In this paper, a direction finding system composed of a small microphone array is used as an intelligent sensor node to lay a foundation for the location of sound source in an intelligent sensor network. The intelligent acoustic sensor network has the advantages of low cost, simple and flexible deployment, high precision and so on. It can be used in outdoor monitoring and control. The direction-finding system in this paper consists of three parts: sensing, processing and communication, which can realize the functions of sound signal acquisition, separation and sound source position direction finding. The intelligent sensor node has a metal encapsulated shell to protect the inner circuit, which can monitor the outdoor environment for a long time instead of manual in the bad environment and weather. In the network structure of intelligent acoustic sensor, considering that the environment monitoring does not need real-time processing, this paper adopts the intermittent real-time processing mode for the monitored sound signal. Limited by the miniaturization of the system structure and the computational performance of the processor, the data processing capacity of the node can not be too large, so it is necessary to complete the computing task within the time limit. Based on the combination of acoustic activity detection (VAD) and generalized cross correlation (GCC), the effective signal segment is intercepted from the collected data on the node processor, and then the direction finding is estimated by using the algorithm PHAT-GCC. The backstage can synthesize the result of direction finding of many nodes and finally realize the joint positioning. By using single node direction finding and multi-node location, the processing capacity on the node and the communication bandwidth on the transmission are reduced. In the process of single node direction finding, due to the influence of the miniaturization of the structure and the encapsulation of the shell, the actual sound source direction finding results obtained in the far field environment have a certain deviation. Based on the above problems, this paper makes a variety of experiments and analyzes the results, and finds out the compensation method for the error, and improves the accuracy of the small intelligent sensor node. It lays the groundwork for the practical orientation of sound source location in intelligent sensor networks.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5;TP212.9

【参考文献】