Robust Multi-robot Object Localization Using Fuzzy Logic Juan Pedro C´ anovas 1 , Kevin LeBlanc 2 , and Alessandro Saffiotti 2 1 Dept. of Information and Comm. Eng., University of Murcia, 30100 Murcia, Spain juanpe@dif.um.es 2 AASS Mobile Robotics Lab, Dpt. Technology, ¨ Orebro University, 70182 ¨ Orebro, Sweden {klc, asaffio}@aass.oru.se Abstract. Cooperative localization of objects is an important challenge in multi-robot systems. We propose a new approach to this problem where we see each robot as an expert which shares unreliable informa- tion about object locations. The information provided by different robots is then combined using fuzzy logic techniques, in order to reach a consen- sus between the robots. This contrasts with most current probabilistic techniques, which average information from different robots in order to obtain a tradeoff, and can thus incur well-known problems when infor- mation is unreliable. In addition, our approach does not assume that the robots have accurate self-localization. Instead, uncertainty in the pose of the sensing robot is propagated to object position estimates. We present experimental results obtained on a team of Sony AIBO robots, where we share information about the location of the ball in the RoboCup domain. 1 Introduction Cooperating robots can benefit in a number of ways from the exchange of infor- mation about perceived objects. For example, a robot which does not directly see an object can still get an estimate of its position. Also, an individual robot’s estimate of an object position can be improved through information sharing. This occurs when some robots have more accurate and/or more reliable po- sition estimates for that object, which can happen due a number of things. For example, a robot may be better localized, have a better view of the ob- ject, or have more effective sensors. In many situations, it is even possible for a group of robots, all having relatively poor estimates of an object’s position, to obtain more accurate and reliable estimates through information sharing. However this requires that the information sharing be performed in an effec- tive way. The problem of cooperative object localization is the problem of fusing infor- mation from different sources in a way which produces agreement about object positions; the agreed upon positions should also be as close as possible to the D. Nardi et al. (Eds.): RoboCup 2004, LNAI 3276, pp. 247–261, 2005. c  Springer-Verlag Berlin Heidelberg 2005