Decision Support Analytic hierarchy process for multi-sensor data fusion based on belief function theory Ahmed Frikha a,b,⇑ , Hela Moalla a,c a LOGIQ Research Group, University of Sfax, Road of Tunis Km 10.5, Technopole of Sfax, BP 1164, 3021 Sfax, Tunisia b Higher Industrial Management Institute of Sfax, Tunisia c Higher Business School of Sfax, Tunisia article info Article history: Received 18 August 2012 Accepted 14 August 2014 Available online xxxx Keywords: Belief function theory Multiple criteria analysis Sensor reading weights Uncertainty measures Conflict measures abstract Multi-sensor data fusion is an evolving technology whereby data from multiple sensor inputs are pro- cessed and combined. The data derived from multiple sensors can, however, be uncertain, imperfect, and conflicting. The present study is undertaken to help contribute to the continuous search for viable approaches to overcome the problems associated with data conflict and imperfection. Sensor readings, represented by belief functions, have to be fused according to their corresponding weights. Previous stud- ies have often estimated the weights of sensor readings based on a single criterion. Mono-criteria approaches for the assessment of sensor reading weights are, however, often unreliable and inadequate for the reflection of reality. Accordingly, this work opts for the use of a multi-criteria decision aid. A mod- ified Analytical Hierarchy Process (AHP) that incorporates several criteria is proposed to determine the weights of a sensor reading set. The approach relies on the automation of pairwise comparisons to elim- inate subjectivity and reduce inconsistency. It assesses the weight of each sensor reading, and fuses the weighed readings obtained using a modified average combination rule. The efficiency of this approach is evaluated in a target recognition context. Several tests, sensitivity analysis, and comparisons with other approaches available in the literature are described. Ó 2014 Elsevier B.V. All rights reserved. 1. Introduction Due to its wide range of applications, multi-sensor data fusion technology has received significant attention in recent research and industry. This technology combines data from multiple sensors to achieve a complete and accurate description of an environment or process of interest. Multi-sensor fusion systems have been widely applied in various areas of robotics, including environment mapping and target recognition, detection and localization. Khaleghi, Khamis, Karray, and Razavi (2013) provided a compre- hensive review for multi-sensor data fusion state of the art, explor- ing its conceptualizations, benefits, challenging aspects, and existing methodologies. The application of multi-sensor data fusion has attracted the attention of several researchers, including Dong and He (2007), Mercier, Cron, Denœux, and Masson (2009) to cite only a few. A single sensor may not be enough to derive a desired level of target estimation or hypothesis identification, and data fusion from multiple sensors is, therefore, often required. It allows to extract a greater volume of information and to attain a more precise level of recognition. Nevertheless, the data derived from multiple sources (signals or humans) is usually imperfect (imprecise, uncertain, and even conflicting). The imperfection and unreliability of sensor data are often attributed to technical and noise (environmental noise, presence of unknown targets, meteorological conditions, etc.) factors. Guo, Shi, and Deng (2006) classified the causes of sen- sor unreliability at three levels, namely the levels of the sensor, the data, and the symbol. Since multiple sensors are uncertain and conflicting, informa- tion fusion becomes a fundamental issue. In fact, most fusion sys- tems are optimistic in that they assume that all sensors are reliable and pay more attention to uncertainty modeling and fusion meth- ods. The performance of the fusion system is, however, highly dependent on sensor performance and adaptability to the working environment and ability to estimate the reliability of each sensor readings (pieces of evidence). Sensor reading reliability needs to be incorporated into the fusion process so as to avoid decreasing system performance (Elouedi, Mellouli, & Smets, 2004; Liu, http://dx.doi.org/10.1016/j.ejor.2014.08.024 0377-2217/Ó 2014 Elsevier B.V. All rights reserved. ⇑ Corresponding author at: LOGIQ Research Group, University of Sfax, Road of Tunis Km 10.5, Technopole of Sfax, BP 1164, 3021 Sfax, Tunisia. Tel.: +216 98 571 500; fax: +216 74 863 092. E-mail addresses: ahmed.frikha@isgis.rnu.tn (A. Frikha), hela_frikha_moalla@ yahoo.fr (H. Moalla). European Journal of Operational Research xxx (2014) xxx–xxx Contents lists available at ScienceDirect European Journal of Operational Research journal homepage: www.elsevier.com/locate/ejor Please cite this article in press as: Frikha, A., & Moalla, H. Analytic hierarchy process for multi-sensor data fusion based on belief function theory. European Journal of Operational Research (2014), http://dx.doi.org/10.1016/j.ejor.2014.08.024