Fuzzy Guidance for Unmanned Ground Vehicles: Theory and Experiments

Abstract

[Abstract] The paper describes the synthesis of a Fuzzy Guidance System for unmanned ground vehicle based on waypoints described in a 4-dimensional space: position in Cartesian coordinates, desired crossing heading, and speed. The vehicle is assumed to be auto piloted in speed, and heading angle. The proposed system uses standard Takagi-Sugeno fuzzy controllers that generate the heading angle reference for the autopilot. A trajectory optimization algorithm is used to yield a long-distance guidance law blended with a shortdistance guidance law as the waypoint approaches. The system handles sets of not directly drivable waypoints, driving the vehicle on trajectories that attempt to cross the waypoints in a prescribed order, with given direction and speed. Extensive computer simulations and outdoor experiments were performed to demonstrate the effectiveness of the proposed method. I. Introduction NMANNED Ground Vehicles (UGV) management is an area of great interest and rapid development, therefore the development of techniques for autonomous navigation in real-world environments constitutes one of the major trends in the current research. The rapid changes in the related technologies, allowed different kinds of mobile robots to be used in many indoor/outdoor applications such as hospitals, offices and factory floors, intelligence service, land maintenance and surveillance, etc. In addition, they offer an ideal test bed for aerospace applications, albeit limited to 2D environment. In the last decade various guidance law were proposed in this regard. Most of the literature in autonomous robotics falls into one of these categories: Trajectory Following, Point-to-Point Motion, or Waypoints Guidance. In trajectory following a reference point on the robot must follow a trajectory in the Cartesian space (i.e., a geometric path with an associated timing law) starting from a given initial configuration; while in point-to-point motion the vehicle must reach a desired goal configuration starting from a given initial configuration. In waypoint guidance the vehicle must reach a set of waypoints with or without a prescribed order starting from a given initial configuration. The present work belongs to the last category, and expands the concept of waypoints guidance introducing desired crossing direction and speed for each waypoint. In a real scenario with bridges or narrow ways, guiding the vehicle across obstacles is very difficult, using traditional guidance laws. Placement of the waypoints requires a large amount of a priori calculation. Starting from this consideration a new guidance law for UGV was studied, implemented and tested. The proposed Fuzzy Guidance System (FGS) assumes that the vehicle knows its current position and next waypoint position. In general, waypoints are described in a four-dimensional space: position in 2-dimensions, desired crossing heading, and speed. For guidance studies, the generic vehicle is assumed to be auto piloted in speed and heading. The FGS is based on standard Takagi-Sugeno