1 Vibroacoustic source separation using an improved cyclic Wiener filter Konstantinos C. Gryllias 1 , Jérôme Antoni 1 and Mario Eltabach 2 1 LVA, INSA-Lyon, University of Lyon, F-69621, France 2 CETIM, F-60300, France {konstantinos.gryllias, jerome.antoni}@insa-lyon.fr, Mario.Eltabach@cetim.fr} Abstract Noise radiated by rotating and reciprocating machines is often a mixture of multiple complex sources, the successful reduction of which is a field of intensive research. In this paper an advanced source separation approach is presented, based on cyclic Wiener filtering, which takes into account the cyclostationarity property of the signals. The aim of the Wiener filter is the separation of noisy measurements into their contributions from the N specific sources and the remaining “noise”. Traditionally this can be achieved by using reference signals which are strongly coherent with the sources of interest and uncorrelated with all the other interfering sources and the masking noise. The Wiener filter can be estimated using the raw signals or only their random part. Moreover, the filter can be underestimated if the Signal-to-Noise Ratio of the reference signals is low, thus leading to the paradox that the level of an extracted source contribution is higher than the overall level. In this study a general strategy is proposed in order to select over which part of the signals (raw or residual) should the filter be estimated. This strategy is based on the number of the available references and the expected number of sources and the link with the multivariable statistical regression. Moreover, in order to increase its robustness, it is proposed to estimate the cyclic Wiener filter using an additional constraint which imposes that the sum of the contributions of the periodic parts of each source equals the overall periodic part as is calculated by the synchronous averaging procedure. This produces a new estimator of the Wiener filter, obtained from a constrained least square optimization. The proposed method is applied on vibroacoustic signals captured on a test rig in order to quantify the contributions of “hydraulic noise” (originating mainly by four hydraulic pumps) and “mechanical noise” (originating from the various rotating parts of the engine). 1 Introduction The interior and exterior noise levels consist a very competitive factor in the market of modern vehicles and machinery. As a result there is increasing demand for developing quitter equipment. Modern machinery is very complex and the noise emitted is finally the combination of several sources. Since the noise produced is of great concern, efforts are being made by the manufactures to develop tools that allow the accurate quantification, separation and prediction of the effects of the sources. The sources are usually both spectrally and temporally overlapping. Therefore two main approaches towards solving the problem have been presented, the separation methods based on a priori knowledge of the noise ([1], [2], [3]) (such as provided by a reference signal) and the separation methods based on the statistical independence of the noise sources (blind source separation methods). These two approaches have been applied intensively to the domain of diesel engines. Antoni et al in [4] and El Badaoui et al in [5] took advance of the characteristics of cyclostationarity in order to perform noise source separation by means of the cyclic Wiener filter. The use of Wiener filter was also explored in [6]. MIMO system modelling was used in [7] in order to estimate the noise transfer function of an engine. On the other hand, in [8] blind source separation methods were used in order to recover signals of different physical sources. This paper is organized as follows. In section 2, the Wiener filter is briefly described. In section 3 a general strategy is proposed in order to select the part of the signals (raw or residual) which should be used