Fuzzy path tracking of autonomous vehicles using GPS

position estimation system based on GPS, that uses a fuzzy sensor fusion technique. A fuzzy path tracking algorithm is also proposed. Both systems have been implemented in the ROMEO-4R vehicle developed at the University of Seville. ÃÃDIUSP9V88DPI Path tracking of autonomous vehicles consists in the generation of the steering commands for a vehicle to follow autonomously a given path taking into account the constraints imposed by the vehicle and the lower level motion controllers. Many path tracking strategies have been presented. In the experiments described in this paper a fuzzy control technique is applied. Position estimation is the process of calculating the position and orientation of the mobile robot with respect to a given reference system. Position estimation is very important for path tracking. If the estimation is not sufficiently accurate, oscillations and erratic behavior of the vehicle may occur. The Global Positioning System (GPS) provides a good source of absolute position information for autonomous navigation, and it can be used for primary position estimation in many cases. However, the receiver can give erroneous measures in certain situations, and therefore, internal sensors as encoders, gyroscopes and compasses have to be used in conjunction with the GPS [4]. The readings of all these sensors are combined using sensor fusion techniques [6]. In this paper, a fuzzy sensor fusion technique has been used. This paper first presents the path tracking problem. Section 3 introduces the Global Positioning System, and its use for mobile robot position estimation. Section 4 presents the fuzzy sensor fusion adopted to deal with the GPS signal degradations. In Section 5, a direct fuzzy controller for the path tracking problem is presented. In Section 6, field experiments with the ROMEO-4R autonomous vehicle are presented. The last three Sections are for the conclusions, acknowledgments and references. !ÃÃQ6UCÃUS68FDIB Several path tracking algorithms select a goal point Q along the path at a fixed distance from the vehicle that is called the lookahead distance G (see Figure 1). Using this goal point three values are obtained that become the inputs to the path tracker. These inputs are the lateral position error in vehicle coordinates (‘,’), the orientation error with respect to the goal point and the curvature error with respect to the goal point. The expressions of and are: (1) where and are the orientation and curvature of the path at the goal point. Direct fuzzy control [1] uses a fuzzy controller …