A parameter optimization approach to the time-minimization of robotic motions along specified paths is presented for the case when: (i) the velocity profile is a prescribed sequence of constant acceleration/deceleration segments with unspecified, but bounded vertex velocities at given path stations; (ii) the relative robot/path location can be varied. Such optimizations occur when technological requirements impose a certain velocity profile along the path due to velocity and acceleration constraints. Full nonlinear manipulator dynamics and path parameterization are used to determine the optimal velocity profile and robot location consistent with the actuator/configuration limitations. No numerical integration or search for switching curve are involved in the solution. Examples of time-and-location optimized robotic motions with specified velocity profile are presented.

The use of ultrasonic range transducers is an inexpensive method of obtaining information about the environment surrounding a robotic System. However, ultrasonic ranging suffers from various shortcomings, one of the major being the production of unreliable data caused by specular reflection.

This work presents an original approach to overcoming the problem of unreliable ultrasonic range data which is caused by specular reflection of the ultrasonic beam. The approach is shown to discriminate between reliable and unreliable data, thus enhancing the role of the ultrasonic range transducer in robotic applications.

This paper presents a new method to generate a smooth collision-free path of a mobile robot. The path is generated as follows: First, the obstacle spaces are artificially contracted to obtain an initial collision-free path which leads the mobile robot from the start point to the goal; then the path is iteratively modified to avoid the obstacle spaces which are gradually restored (this is carried out in consideration of the index introduced to evaluate the correctness of the path), in the final step, all the obstacle spaces are fully restored and the obtained path becomes a desirable collision-free one. The proposed method is effectively applied to an example of planning planar movements of a mobile robot.

Intelligent control of mobile systems allows for hierarchical structures that utilize sensory data with various levels of accuracy. This paper discusses a rule-based approach for the control problem. The assumed inexactness in world description is represented by fuzzy memberships, and the state space is discretized into a linguistic vocabulary. Fuzzy motion control rules that have been experimentally derived, are then used in a fuzzy inference mechanism to give the final control command to robot actuators. Finally, the developed algorithm is tested for real-time control applications.

A practical architecture, using a four-bar-linkage, is considered for the University of Minnesota direct drive rotot. This statically-balanced direct drive robot has been constructed for stability analysis of the robot in constrained maneuvers.2–6 As a result of the elimination of the gravity forces (without any counter weights), smaller actuators and consequently smaller amplifiers were chosen. The motors yield acceleration of 5 g at the robot end point without overheating. High torque, low speed, brush-less AC synchronous motors are used to power the robot. Graphite-epoxy composite material is used for the construction of the robot links. A 4-node parallel processor has been used to control the robot. The dynamic tracking accuracy-with the feedforward torque method as a control law- has been derived experimentally.

A novel approach to control synthesis of biped locomotion mechanisms is suggested. The synthesis is carried out in two stages: the stage of nominal regimes (the synthesized control has to ensure the realization of gait in the absence of any disturbance), and the stage of perturbed regimes (the control has to eliminate deviations from the nominals under an additional condition of preserving the stability of the overall System). At the level of perturbed regimes, the proposed control synthesis should ensure the compensating movements such to bring the System ot its nominal during the gait. However, the compensating movements can, as a side–effect, induce the undesirable inertial forces which can influence the mechanism overall stability. To avoid this, it is suggested such a control which ensures that the acceleration of compensating movements does not exceed a value given in advance. In addition, the case is considered when an additional correction of the zero moment point ZMP position is accomplished by different mechanism joints.

This paper presents a linearization by static feedbacks in the robotic field, i.e. by feedback depending on the whole state space. A phenomenological approach is considered, which by using the derivation with respect to time, leads to the major results of the method. Simulation results are presented, and some aspects of the correction of the effect of the characteristic numbers are also discussed.

A method for planning minimum time joint trajectories for robot manipulators is discussed. The minimum time trajectory planning problem for manipulators is one of the minimum time control problems of non-linear systems. The optimal input torque/force is of a bang-bang type, except for the singular control derived from the Maximum Principle. An algorithm for generating a bang-bang control is proposed. In the algorithm, the switching time vector is updated to decrease the final state error. The proposed algorithm is applied to a simple manipulator with two links, and the solution by this algorithm is compared with the sub-optimal solution obtained by another approximate method.

The Autonomous Robot Architecture (AuRA) provides multi-level representation and planning capabilities. This paper addresses the task of navigational path-planning, which provides the robot with a path guaranteed to be free of collisions with any modeled obstacles. Knowledge supporting visual perception can also be embedded, facilitating the actual path traversal by the vehicle.

A multi-level representation and architecture to support multi-sensor navigation (predominantly visual) are described. A hybrid vertex-graph free-space representation based upon the decomposition of free space into convex regions capable for use in both indoor and limited outdoor navigation is discussed. This “meadow map” is produced via the recursive decomposition of the initial bounding area of traversability and its associated modeled obstacles. Of particular interest is the ability to handle diverse terrain types (sidewalks, grass, gravel, etc.) “Transition zones” ease the passage of the robot from one terrain type to another.

The navigational planner that utilizes the data available in the above representational scheme is described. An A* search algorithm incorporates appropriate cost functions for multi-terrain navigation. Consideration is given to just what constitutes an “optimal” path in this context.

This paper presents a particular architecture of a parallel robot which is characterized by a diagonal stiffness matrix. This result suggests the use of a parallel robot in the final phase of insertion as a passive compliance device. The stiffness rate of this device is controlled by the gain of the feedback loop. As the correction of small angular misalignments due to contact forces do not generate lateral errors and vice versa, we have the equivalent of an isoelastic swivel.

In this paper the authors present a mathematical model of geometric and kinematic behaviour of an original passive compliant device provided with two rotation centres, called DCR–LAI system. This device is designed for a robotic assembly of parts with very small tolerances including a chamfer at the hole. The given modélisation may be used as a decision aid for the choice of a compliant device with regard to characteristics of parts being assembled.

Solving the direct kinematic problem in a symbolic form requires a laborious process of successive multiplications of the link homogeneous transformation matrices and involves a series of algebraic and trigonometric simplifications. The manual production of such solutions is tedious and error-prone. Due to the efficiency of the Prolog language in symbolic processing, a rule–based Prolog program is developed to automate the creation of the following processes: Link transformation matrices; forward kinematic solutions; and the Jacobian matrix. This paper presents the backward recursive formulation techniques, the trigonometric identity rules, and some heuristic rules for implementing the System. A verification of the System has been demonstrated in case of several industrial robots.

An extension of the inverse kinematics algorithm by Gupta and Kazerounian is presented. The robot kinematics is formulated by using the Zero Reference Position Method. Euler parameters and the related vector forms of the spatial rotation concatenation have been used to improve the efficiency of the velocity Jacobian computation. The joint rates are formally integrated by using a modified predictor-corrector method particularized to robot inverse kinematics – it is a strict descending, p(1)c(0 – n), variable step algorithm. The definitions of the rotational error and overall error measure have been revised. Depending upon the convergence criteria used, these modifications reduce the overall computational time by 20%.

The author introduced the use of time–varying inertia links to increase the flexibility (or versatility) of manipulators in the sense that they can so achieve more tracking capabilities. Clearly, we then have a pair of control vectors (link inertias and external forces) instead of one control vector only (applied forces). After a careful derivation of the dynamical equations of such manipulators, one shows that their control procedure can be decomposed into two stages, each of them involving sliding control schemes via prescribed dynamics on the error term, that is to say the difference between the desired position and the actual position. This approach provides a good technique to take account of possible constraints on the magnitude of the control, and in addition, it is possible to consider the sliding conditions as an ideal mechanical constraint, therefore the use of the principle of virtual displacements.

This paper is a continuation of the one entitled “Trajector control of manipulators with time–varying inertia links” (Robotica 6, No. 3, 197–202, 1988). It described new techniques to circumvent difficulties arising when the values of the mechanical parameters are not exactly known and when linearization does not apply because of the important deviation. In addition this paper shows that Appel's principle of virtual displacements is more efficient than the Lagrangian parameters for the derivation of the manipulator's dynamic parameters in the presence of mechanical constraints.

An off-line programming station for a robotised welding cell is presented. In this context the choice of a kinematic inversion method applied to the simulation of robots is discussed.

A direct inversion method is developed in the case of 5-axes robots of plane geometry and for a task presenting a symmetry of revolution. It uses the geometry of the tool and of the robot and vectorial calculations very well adapted to this particular case. The results presented validate the choice of this method for a specialised programming environment.

Accurate modeling of robot dynamics is a prerequisite for the design of model-based control schemes and enhancement of the performance of the robot. The dynamic parameters associated with a pseudo-inertia matrix are often difficult to identify accurately because the inertia torques are small in comparison to gravity loadings, thus creating signal processing problem. The identification method presented in this paper utilizes a balancing mechanism which increases the estimation accuracy of the dynamic parameters. The balancing mechanism has the effect of amplifying the inertia-related torque signal by eliminating gravity loadings acting on the robot joints. A series of motion data were experimentally obtained through sequential test steps. By incorporating the measured information about joint torques, angular positions, velocities and accelerations the least square algorithm was used to identify the dynamic parameters. The estimated values were converted to those of the original robot model to obtain its dynamic model parameters. The identified robot dynamic model was shown to be accurate enough to predict the actual robot motions.

A method was developed that takes into account flexibility of robot links in the inverse dynamics calculations. This method uses the Newton-Euler equations and is applicable for special case systems that allow for only a small degree of flexibility. Application of the method should improve the accuracy of the position of the end effector during motion of the robot.

The results of this study show that the method can be based entirely on an existing rigid-link model with only minimal changes required as additions. The computational complexity of the method is discussed briefly as well and indicates an increase of computations of slightly more than a factor of two as compared to a rigid-link model for the same robot geometry.

Studies are the effects of joint flexibility on the dynamic stability of a one-link force-controlled manipulator. The closed-loop dynamic equation with the explicit force controller and the damping force controller are first derived. Stability analysis is then carried out by computing the system eigenvalues. Results indicate a conditionally stable system. Due to the presence of discontinuous contacts with the environment during the interaction process, the system exhibits a stable limit cycle when the force feedback gain goes beyond the critical value.

The development of robot languages has followed a pattern similar to that of conventional programming languages, where robot languages have been based on an existing programming language. This paper first identifies the use of an existing base as one way of developing robot programming languages, and discusses the areas of difficulty in this approach. Then, on-line and off-line programming of robots is discussed and the requirements of robot programming languages that are different to those of non-specialised programming languages are presented. A discussion and evaluation of some programming languages in terms of their appropriateness for use as the base for an intelligent robot programming language is presented. This leads to the conclusion that no current language forms an adequate base for intelligent robot programming languages. What is needed as a base is a language for use in the artificial intelligence domain, that incorporates real-time facilities.