ELSEVIER Fuzzy Sets and Systems 69 (1995) 15 27 FUZZY sets and systems Fuzzy logic and autonomous vehicles: Experiments in ultrasonic vision M. Poloni*, G. Ulivi, M. Vendittelli Dipartimento di ln/ormatica e Sistemistica. Uniuersitlt degli Studi di Roma "'La Sapienza", Italy Received September 1993:revised March 1994 Abstract The opportunities offered by fuzzy logic to build maps for robot navigation are investigated. Characteristics of points of the space (occupied, free, uncertain, etc.) are easily expressed through set theoretical operations. Real-world experi- ments validate the approach. The experimental set-up is based on modified Polaroid ultrasonic sensors; however, the approach can be easily extended to incorporate other kinds of sensors. Keywords. Data analysis methods; Linguistic modeling; Robotics 1. Introduction Use of mobile robots is spreading in many fields of application. Typical examples are submarine and spatial exploration [21, 11, 9], surveillance of industrial installations, rescue missions in dangerous environ- ments (radiations, chemical contamination). More common applications are, for example, internal mail delivery, partially autonomous wheel-chairs for disabled [-20], intelligent Automatic Guidance Vehicles (AGVs) for floor-shops [2]. Despite the differences among these applications, an autonomous vehicle always needs a navigation system to determine a suitable path to its target and a sensory system to acquire knowledge about the crossed environment. Consider the simplified functional structure of a navigation system depicted in Fig. 1. At the beginning the map can be totally or partially unknown. The lowest level is constituted by the measuring processes that supply rough information coming from each sensor, typically the distance of the nearest obstacle (Sensors). Information is suitably merged and used to build or upgrade the map of the traveled environment (Map reconstruction). The map is used to implement the navigation strategy. It can also be interpreted by other levels of the structure, trying to identify the shape and possibly the nature of the objects on the scene (Map interpretation) to devise more complex strategies [18]. In this case the map could contain other details on the environment as, for example, the colors of the obstacles. *Corresponding author. Correspondence address: ELITE Foundation, Promenade 9. 52076 Aachen, Germany. 0165-0114/95/$09.50 CO1995 Elsevier Science B.V. All rights reserved SSDI 0165-0114(94)00237-1