SENSOR INTEGRATION ON A MOBILE ROBOT Geert De Cubber Hichem Sahli Francis Decroos VRIJE UNIVERSITEIT BRUSSEL – ETRO DEPARTMENT Pleinlaan 2 – 1050 Brussel – Belgium {gdcubber, hsahli, frdecroo}@etro.vub.ac.be ABSTRACT The purpose of this paper is to show an application of path planning for a mobile pneumatic robot. The robot is capable of searching for a specific target in the scene and navigating towards it, in an a priori unknown environment. To accomplish this task, the robot uses a colour pan-tilt camera and two ultrasonic sensors. As the camera is only used for target tracking, the robot is left with very incomplete sensor data with a high degree of uncertainty. To counter this, a fuzzy logic - based sensor fusion procedure is set up to aid the map building process in constructing a reliable environmental model. The significance of this work is that it shows that the use of fuzzy logic based fusion and potential field navigation can achieve good results for path planning. 1. INTRODUCTION Mobile robots have attracted several research activities for many applications, such as navigation. The key research areas are: • Sensor technology: Mobile robots can be equipped with a variety of sensors, enabling the robot to gain some knowledge about its surroundings. The most common sensors are the infrared and ultrasonic devices, used in al sorts of configurations [2][3][4]. On the other hand, computer vision can be applied for robot navigation, using either monocular or stereo vision [5][6][7][8]. No matter what sensors are used, the basic problems to be solved are always the same: o What measuring strategy should be applied in order to collect the maximum amount of data in the minimum amount of time? o How can the error on the readings be minimized? o How big is the measurement error? • Sensor fusion: If a robot needs to gain a more or less complete “image” of its environment, it cannot rely on only one type of sensor. Hence the need for an intelligent sensor fusion algorithm to combine the often erratic, incomplete and conflicting readings received by the different sensors, to form a reliable model of the surroundings. Sensor fusion has been subject to a lot of research [12][13], most of the proposed methods use Kalman Filtering [17] and Bayesian reasoning [15]. However, in recent years, there has been a tendency to make more and more use of soft computing techniques such as artificial neural networks [14] and fuzzy logic for dealing with sensor fusion. [16]. • Map building & path planning An autonomous robot must keep a kind of map as a model of its surroundings. These maps can be simple grid maps, topological maps [19], or integrated methods [20]. The used path planning technique depends highly upon the type of map chosen before. A survey of different methods can be found in [9]. • Efficient control strategies: All the different processes (sensor measurements, measurement processing, sensor fusion, map building, path planning, task execution …) must be coordinated in an efficient way in order to allow accomplish a higher goal [21]. A number of control strategies can be set up, varying from simple serial sense-model-plan-act strategies to complex hybrid methods. A discussion of some of these control strategies can be found in [22]. An interesting approach here, is to use fuzzy behaviours, partially overriding each other, to build up complex navigation plans, as discussed in [23][24][25]. In this paper we present a hybrid control strategy. The rest of this paper is organized as follows: The platform and control strategy are described in section 2, the data fusion and modelling are summarized