噪声源识别及分离技术是内燃机噪声研究的一个重要领域，本项目提出了一种基于计算听觉场景分析的新的内燃机噪声源识别方法，研究将立足于借鉴对人听觉感知的研究成果建立一系列的模型和算法，并通过软件编程来实现，通过内燃机的噪声试验及分离来验证方法的有效性，提出评价分离效果的定量指标。主要研究内容包括：进行内燃机噪声信号的不同获取方式的试验，确定最佳的获取点和测点的组合方式，为噪声源信号分离算法提供包含充分源信息的混合信号，为分离算法的检验提供充分的信号库；应用理论分析和仿真来研究噪声信号的变换、分离与组织的算法，研究整套算法的软件实现，通过标准声音信号库和实测的噪声信号库来验证算法的有效性，并进一步改进与修正数学模型及算法。通过实验研究内燃机独立噪声源的获取方法，并据此提出评价分离效果的定量指标，本项目研究将为实际应用中分析各种源的特征和贡献量以及内燃机的声音品质设计奠定理论基础。

Noise source identification and separation technology is an important research field of engine noise control. A new computational auditory scene analysis based method is proposed to separate IC engine noise signal. The aim of this project is to build a series of models and algorithms based on researches about human auditory perception. Realized by software, validity of the method can be checked through the IC engine noise experiments and the noise separation tests. Then quantitative index for evaluation of separating effects is studied. Main researches include: a series of engine noise measuring experiments are carried out to find good measuring positions and sensor number. After that,experiments of different working conditions are done to get enough noise signals with adequate information for verifying separation algorithms. By theoretical analysis and simulation, algorithms about signal preprocessing, separation and grouping are studied. Then the implementation of these algorithms is carried out and software is programmed using MATLAB. By the above mentioned researches,a noise signal separation system is got. The effectiveness of the system is tested with standard voice signals and the measured IC engine noise signals. Following this, further improvements and amendments of the mathematical models and algorithms are achieved. By designing special experimental equipment, independent noise source measurment is explored. Based on this experimental research, quantitative index about accuracy of speration is derived. The research of this project can provide accurate information for the analysis of characteristics of noise sources and their contribution to IC engine noise level. Finally the method and the system can be used for the sound optimization of various internal combustion engine.