Navigation-guidance-based robot trajectory planning for interception of moving objects

grantor: University of TorontoMotivated by current trends in automation of industrial applications, the general problem area addressed in this thesis is "on-line robot-motion planning for intercepting randomly moving objects." The specific objective of the thesis is "the use of navigation-guidance techniques in the development of a generalized scheme that can intercept a randomly moving object with time-optimality". Two novel methods are presented for on-line-robotic-interception of fast-maneuvering objects which do not require a priori information on the moving-object's motion. Both techniques combine a navigation-guidance-based method with a conventional object-tracking technique. Thus, they are classified as 'hybrid' interception schemes with two phases: Phase I during which the robot is under the control of a navigation-guidance-based technique, and Phase II during which the robot's control is switched to a conventional tracking method. For the first proposed method, an Ideal Proportional Navigation Guidance (IPNG) technique is used during Phase I. This technique moves the interceptor rapidly toward the rendezvous point. For the second proposed method, the augmented form of the IPNG technique is suggested for Phase I, when a reliable estimation of the target's acceleration can be provided to the interceptor. Both techniques are modified in this thesis to reflect the greater mobility of a robotic manipulator over an airborne missile for robotic interception of fast-maneuvering targets. Since IPNG techniques have been originally designed for missile guidance, they do not attempt to match the target's velocity at the interception point. In this thesis, for smooth interception, a tracking method is proposed to be switched on to at an optimal time in order to bring the robot to the interception point matching both target's position and velocity. On-line selection of this time-optimal switching point is discussed in the dissertation. The convergence of the proposed interception methods under ideal conditions is addressed. The effect of noise in target's position readings on the on-line estimation of the interception time, and, subsequently, on the overall performance of the proposed technique is also discussed. Extensive computer simulations illustrate the effectiveness of the IPNG-based interception methods developed in this thesis over pure tracking-based techniques proposed in the cited literature.Ph.D