This paper discusses one problem of robot dynamics rarely mentioned in papers relevent to this field. It is the problem of torsional effects in torque transmissions (reducers, shafts, transmission chains, etc.). The problem is significant since oscillations can appear to be due to these effects. The complete dynamic model, which includes these effects, is derived and the possible simplifications considered. The position of feedback transducers is discussed since it appears as an important problem when it is intended to minimize the influence of these elastic vibrations. The discussion is based on eigenvalues and simulation results.

Fluid power is one of the vast fields and its applications in manufacturing are not used for many real-time problems though its capabilities are known. One such application is fluid power for robotic systems. More specifically, in this study, a new hydraulic system is developed for use in actuating a robotic gripper. Electric motors were found unsatisfactory due to their poor power to weight ratio. Hydraulics were attempted using a master slave hydraulic system. Powering and positioning was done on the master cylinder, while the remote slave cylinder followed the motion. The control of the master cylinder was done with a ground based stepping motor and lead screw mechanism. Small hoses ran to the gripper based slave cylinder, which copied the motions.

In this paper, we develop a method for mobile robot control using ultrasonic sensors to determine displacement constraints between two parallel planes. Robot control has been simulated for this situation. Results obtained by simulation and experimentation are presented. A degraded mode which considers the use of one valid ultrasonic sensor is studied. In this case, a new model is developed and new results are analyzed.

The paper addresses the following problem: for a three-bar lifting linkage, find the locations of the two ends of the actuated link so as to minimize the maximum value of the force seen by the actuator over a specified range of arm angular positions. A simple analytical solution to this min-max problem is given, showing that the optimal locations are with the actuated link perpendicular either to the fixed link or to the lifting arm, when the arm is in a horizontal position. The method can be used to optimize the actuation of robots, doors, platforms, landing gears, etc., and allows to account for actuator length constraints. It can be extended also to Stewart's platform-type spatial manipulators.

An efficient on-line scheme for computing the inverse joint solution of robotic manipulators is combined with an improved formulation of robust, non-linear feedback control in joint space to produce a realizable Cartesian control scheme. Parametric uncertainties in the robot model are highlighted by the inclusion of compliance at each joint. Simulation results for a two link, coupled manipulator demonstrate that this Cartesian control enables the tip of the arm to track the demanded trajectory with arbitrarily small error in response to realistic actuator torques.

The need for fault tolerant mechanisms in flexible manufacturing systems is described and previous work on diagnosis in robotics and other areas is considered. Fundamental difficulties in the analysis of robot cell malfunctions are described and a glossary of terms useful in this area is presented. Limited observational data on the occurrence of faults in assemblies are reported. Finally a proposal for an experimental mechanism for diagnosis within a knowledge rich supervisory system is explored.

In this paper, we discuss the accurate and robust sliding mode tracking control for highly nonlinear robot manipulators using a disturbance observer. Due to the modeling error or environmental uncertainties, sliding mode control may present a significant chattering problem by using a conventional Sliding Mode Control (SMC) which can obtain the desired tracking performance, because the controller design is carried in the uncertainty space of the system parameters.

To solve this chattering problem, the efficient compensation of the disturbance observer has been introduced. Thus the design of the proposed SMC is not influenced by the modeling error or parameter uncertainties.

To ensure bounded stability, the proposed sliding mode controls have been analysed in a preliminary lemma and two main theorems. The stability of the proposed control method is proved in this paper, and the efficiency of the control algorithms has been demonstrated by simulations for a position tracking control of a two-link robot subject to parameter and payload uncertainties.

A new mathematical formulation of robot and obstacles is presented such that for on-line collision recognition only robot joint positions in workspace are required. This reduces calculation time essentially because joint positions in workspace can be computed every time from the joint variable through robot geometry. It is supposed that the obstacles in the workspace of the manipulator are represented by convex polygons. For every link of the redundant robot and every obstacle a boundary ellipse in 2D is defined in workspace such that there is no collision if the robot joints are outside these ellipses. First, some methods are presented for the automatic determination of these ellipse functions from the obstacle and robot data. Then, the boundary ellipse functions are used as optimization criterion in a collision-avoidance method. The method permits the tip of the hand to follow a given path in the free space while the kinematic control algorithm maximizes the boundary ellipse function of the critical link. The effectiveness of the proposed methods is discussed by theoretical considerations and illustrated by simulations of the motion of three- and four-link manipulators between obstacles.