Object Detection by an Autonomous Robot through Data Fusion Tapan Kumar Nayak #1 , Sanjit Kumar Dash *2 , Himansu Sekhar Sasmal #3, Srikanta Patanaik #4 1, 2 College of Engineering and Technology Biju Patanaik University of Technology Bhubaneswar, Odisha, India 3, 4 Institute of Technical Education and Research Siksha “O” Anusandhan University Bhubaneswar, Odisha, India Abstract— Wireless sensor network is an emerging technology that enables remote monitoring objects and environment. Object tracking is an important wireless sensor networks. Single sensor system cannot provide satisfactory accuracy or reliability as a result of which a system deploys multiple sensors in a particular scenario. To handle uncertain, incomplete and vague information, data fusion technique is applied to a measuring system to reduce effectively overall uncertainty and increase the accuracy. In data fusion Dempster-Shafer evidence theory is an efficient method. Here different bodies of evidences are combined by using Dempster’s combination rule. This paper gives an approach to find the exact location of goal object through data acquisition and fusion by deploying multiple sensors. The work is concentrated on robot how to track the targets and obtain the target position effectively. Keywords— Mobile Robot, Dempster Shafer Evidence Theory, Data Fusion I. INTRODUCTION The processing of incomplete, uncertain and vague information is required to get the best approximation. To deal with these factors different approaches like Bayesian probability theory, fuzzy logic and Dempster_Shafer Theory are there .The Bayesian theory requires complete prior knowledge of probability of evidences .As we don’t have any prior knowledge we cannot use Bayesian Theory here. Fuzzy logic can be applied to combine evidences but fuzzy requires the complete knowledge of membership functions for fuzzy set which is not easy to obtain in real world application. We choose DS Theory because it supports the representation of both imprecise and uncertainty and it is able to deal with ignorance and missing information. Information fusion technique was introduced to get more accurate and integrated data to real world application. Multi-sensor system combines the information and gives more accurate and integrated data than a single sensor. Dempster_Shafer theory allows information integration by both belief and disbelief. Information fusion is central to many computer systems that help users in decision making, forecasting, pattern recognition and diagnosis. A complicating factor in developing these decision support systems is the handling of the uncertainty resulting from imprecise, incomplete, vague and complementary information. The aim of the management of uncertainty in these systems is to get the best approximation. The main basic approaches to uncertain reasoning are certainty factors developed in expert systems, Bayesian probability theory, fuzzy logic and Dempster- Shafer (D-S) evidence theory[1-6].The Bayesian approach has a decision making theory, but it requires complete knowledge of combined conditional probabilities and specification of the prior knowledge of probability distribution proving that a piece of evidence is present. Besides, the main limitation of the Bayesian approach is that it cannot model imprecision. That is, the Bayesian probability theory cannot measure a body of evidence with an imprecision on probability measurement. The main advantage of the fuzzy fusion approach is that the evidence from multiple features can be combined using fuzzy logic operations, and uncertainty can be represented. The fuzzy set framework provides a lot of combination operators, which allows the user to adopt a fusion scheme and specify the data at hand. However, to our knowledge, the membership functions for the fuzzy set are not easy to obtain in real-world application systems. There are three main reasons why the D-S evidence theory should be taken into account when it comes to information fusion. First of all, since the D-S evidence theory supports the representation of both imprecision and uncertainty, it is considered to be a more flexible and general approach than the traditional probability theory. Secondly, D-S offers the possibility of coming up with the probabilities of a collection of hypotheses, whereas a classical probability theory only deals with one single hypothesis. Finally, the major strength of the D-S theory is its ability to deal with ignorance and missing information. The most crucial rule of evidence combination in the D-S theory is called the Dempster’s rule of combination. It has several interesting mathematical properties [3]. However, combination may yield illogical results when the collected bodies of evidences highly conflict with each other. The conflicting evidence has been one of the most important issues in D-S evidence theory [7]. Many methods have been proposed to solve this problem, but there has been no solution accepted universally so far. Information fusion technology not only eliminates redundancy but also obtains more accurate and integrated estimation than from any single source. Dempster-Shafer theory is a kind of reasoning algorithm based on evidence theory [3]. It was Tapan Kumar Nayak et al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 3 (5) , 2012,4951-4955 4951