Engine fault diagnosis based on multi-sensor information fusion using Dempster–Shafer evidence theory q Otman Basir * , Xiaohong Yuan Department of Electrical and Computer Engineering, University of Waterloo, 200 University Avenue, Waterloo, Ont., Canada N2L 3G1 Received 26 February 2004; received in revised form 12 July 2005; accepted 12 July 2005 Available online 25 October 2005 Abstract Engine diagnostics is a typical multi-sensor fusion problem. It involves the use of multi-sensor information such as vibration, sound, pressure and temperature, to detect and identify engine faults. From the viewpoint of evidence theory, information obtained from each sensor can be considered as a piece of evidence, and as such, multi-sensor based engine diagnosis can be viewed as a prob- lem of evidence fusion. In this paper we investigate the use of Dempster–Shafer evidence theory as a tool for modeling and fusing multi-sensory pieces of evidence pertinent to engine quality. We present a preliminary review of Evidence Theory and explain how the multi-sensor engine diagnosis problem can be framed in the context of this theory, in terms of faults frame of discernment, mass functions and the rule for combining pieces of evidence. We introduce two new methods for enhancing the effectiveness of mass functions in modeling and combining pieces of evidence. Furthermore, we propose a rule for making rational decisions with respect to engine quality, and present a criterion to evaluate the performance of the proposed information fusion system. Finally, we report a case study to demonstrate the efficacy of this system in dealing with imprecise information cues and conflicts that may arise among the sensors. Ó 2005 Published by Elsevier B.V. Keywords: Evidence theory; Engine diagnosis; Information fusion; Sensor fusion; Fault detection and identification; Pattern recognition 1. Introduction Information fusion, when applied to fault diagnosis and defect inspection, revolves around two main ques- tions [1–3]: (1) How to acquire precise and reliable information cues about potential faults by incorporating com- plementary, and possibly redundant, multiple sensors. (2) How to fuse decisions that are derived based on multi-sensor data, which can be imprecise, and conflicting. In the context of engine diagnosis, the first question is concerned with extracting fault-reveling engine features from multiple sensors, and describing them in a coherent representation scheme. Furthermore, since information obtained from the sensors is inherently incomplete, uncertain, and imprecise, it is imperative that a fusion mechanism be devised so as to minimize such impreci- sion and uncertainty. The effectiveness of such mecha- nism depends to a large extent on how redundant and complementary are the information cues obtained from the sensors. It is equally important to decide at what level of abstraction the fusion process is to take place, e.g., at the measurement level, at the feature level, 1566-2535/$ - see front matter Ó 2005 Published by Elsevier B.V. doi:10.1016/j.inffus.2005.07.003 q Editorial note: Dr. Zheng Liu, Institute for Aerospace Research, National Research Council, Canada, assisted in managing the review process for this manuscript. * Corresponding author. Tel.: +1 519 885 1211x6754; fax: +1 519 7459774. E-mail address: obasir@uwaterloo.ca (O. Basir). www.elsevier.com/locate/inffus Information Fusion 8 (2007) 379–386