多源运动声场声特征提取和空间稀疏化方法及测量关键技术研究

The ground of road traffic noise is one of the main source of pollution city environment. For this type of noise has multi-sound sources, wide radiation, acoustic features and motion characteristics of coupling, the measurement and monitoring become difficult problem, and its relevant acoustic theory basic research should be carried out as soon as possible. Therefore, the project intends to multi-source movement sound source as object to carry out the following research: (1) Investigating multi-source movement characteristics of sound radiation and transmission of sound field as well as its extraction method, analyzing the acoustic radiation characteristics, sound propagation characteristic and its coupling with motion state, establish 3D model of the sound radiation and sound transmission based on the movement characteristics of sound; (2) Study multi-source movement sound radiation space sparse methods based on the characteristics, establish a sound source with multi-source movement sound field, the sound propagation model and its coupling with movement, build sparse acoustic characteristics of multi-source movement sound field and the sound radiation model on the bases of sparse acoustic features; (3) Research multiple source sound field area noise measuring random acoustic array and monitoring technology, set up area noise monitoring system. This project is expected to reveal the acoustic radiation and sound transmission characteristics of multi-source movement sound field, realizing sparse acoustic expression of the acoustic characteristics, building measuring and monitoring method and technology of multi-source movement sound field area. Research results will provide the scientific basis for noise source identification, noise measurement and monitoring of moving noise field.

地面道路交通噪声是影响城市环境的主要污染源之一，这类噪声以声源多、辐射面大、声特征与声源运动状态耦合性强而成为噪声测量和监测的难题，其相关声学理论基础研究亟待开展。为此，本项目拟以多源运动声场为对象开展以下研究：一是，研究多源运动声场的声辐射和声传播特征及其提取方法，分析多声源声辐射特征、声传播特性与声源运动状态的耦合性；二是，研究基于声特征的多源运动声空间稀疏化方法，确定多个运动声源声传播模式及其与运动的耦合关系，建立多源运动声场的稀疏化声特征和融合声特征的稀疏化声辐射模型；三是，研究基于稀疏化方法的测量随机声阵列搭建和测试技术，构建多源运动声场的监测系统。项目研究将确定多源运动声场的声辐射和声传播特性，给出声特性的稀疏化表达和多源运动声场重构方法，开发运动噪声声场的监测技术。研究成果将为运动噪声源识别、噪声场测量与监测提供科学依据。

随着我国交通运输业高速发展，交通道路扩建、改造、新建速度进一步提高，由此引起的交通噪声环境污染日益严重。此类噪声场是一种典型的多源运动声场，特征复杂、重建难，可视化要求高，而多源运动声场特征及测量是进行道路交通噪声治理的前提和基础。为此，本项目从多源运动声场的声特征提取、声场稀疏化及其重建、声场稀疏化测量三方面开展了研究。主要研究内容包括：声特征提取方法、声场稀疏化机理及稀疏重构方法、声场稀疏测量及声阵列构型。经历时4年的研究，取得了如下重要科研成果：首先，从多源运动声场的声特征入手，针对非嘈杂和嘈杂声环境两中情况下多源运动声场的声特征提取进行研究，提出了以流形学习算法为主的STFT-LLE和SVD-LLE的两种流形学习方法，建立了不同背景声影响下多源运动声场的声特征提取方法及实现流程，通过数学解及边界元数值仿真预测不同类型运动声源的声场特性，同时考虑多个运动声源的不同声特征间的重叠现象，实现了多声源声场的低维流形物理意义与声场特征间一一对应性解释，拓展了流形学习算法的应用空间。其次，基于声场的特征及空间分布稀疏性，在多源运动声场的空间稀疏化、声空间稀疏重构上取得关键方法和技术突破。实现了利用多源运动声场的稀疏插值重建，为声场的稀疏化测量奠定理论及技术基础。最后，依据上述理论和技术研究结果与道路交通噪声特点，建立了完整、实用的道路交通噪声场在线监测、预估和评价的多源运动声场稀疏化重构模型，并搭建了融合声特征道路交通环境噪声的稀疏声阵列测量系统。该项目研究成果不仅提出了多源运动声场的声特征提取方法、稀疏化重构和稀疏化测量方法，而且实现了多运动目标实时定位和多声特征信息融合的声阵列测量技术关键突破。解决了交通噪声测量和重建难，预估无依据不准确和监测无法实时可视的难题，同时为实现道路交通运动噪声场的监测、预估、评价提供理论方法依据，为后续运动声辐射理论和阵列测量技术进一步研究奠定了坚实基础。