Experiments of Fuzzy Real-Time Path Planning for Unicycle-Like Mobile Robots under Kinematic Constraints

This paper presents an experimental study concerning the application of a real-time motion planning algorithm to an unicycle-like mobile robot. The desired trajectory to be followed by the mobile robot in presence of bounds on the linear/angular velocities and accelerations is available on-line. Moreover, in the respect of the kinematic constraints, the desired path has to be kept as long as possible. The implemented algorithm is based on a discrete-time kinematic control which implement a warping of the time law based on the definition o / a virtual time. In addition, a fuzzy inference system handles the additional information given by the difference between the virtual and real time in order to exploit knowledge in advance of the desired path. The experimental results confirm the effectiveness o/ the adopted algorithm. 1 I n t r o d u c t i o n In several applications the problem of steering a mobile robot with unicycle-like kinematics arise. The vehicle is usually subject to kinematic constraints such as bounded linear and angular velocities and accelerations; moreover path tracking capabilities with a the desired trajectory available in real-time are often required. References [1, 3, 4] deal with the motion control problem for the end effector of a robot whose trajectory is given in real-time; reference [3], however, does not take into account acceleration constraints that are included in references [1, 4]. In a previous paper [2] a motion planning algorithm for an unicycle-like mobile robot in presence of velocity and acceleration constraints has been presented. This algorithm satisfies the kinematic constraints, given in terms of linear and angular maximum velocities, and dynamic constraints, often given in terms of maximum linear and angular acceleration e.g., as in [8], while keeping the desired path. The algorithm is based on a discrete time kinematic control which implements a warping of the time law that makes use of a virtual time. Moreover, a fuzzy inference system handles the additional information given by the difference between virtual and real time when the virtual time is slowed down. After briefly recalling the main steps of the algorithm, this paper presents an experimental study concerning the application of the considered motion planning algorithm to the Magellan Pro mobile robot manufactured by Real World Interface [7]. The experimental results show the behavior of the considered robot and validate the effectiveness of the adopted procedure. 2 P r o b l e m f o r m u l a t i o n