A modular hierarchical model for controlling robots is presented. This model is targeted mainly for research and development; it enables researchers to concentrate on a certain specific task of robotics, while using existing building blocks for the rest of their applications. The presentation begins by discussing the problems with which researchers and engineers of robotics are faced whenever trying to use existing commercial robots. Based on this discussion we propose a new general model for robot control to be referred as TERM (TEchnion Robotic Model). The viability of the new model is demonstrated by implementing a general purpose robot controller.

The previous paper by the same authors (Robotica 2, Part 2, pp. 93–103 1984) was concerned with the application of logic programming to the generation of paths for robots. This paper is dealing with the PIGS system for robot problem-solving in generating plans for robots. The system avoids the overhead of generating a new set of facts for each state considered. A computer-type control is adopted not only to generate a simple sequence of actions to be performed by the robot, but also to enable the robot to follow an accurately defined path. A suitable plan of action of a robot problem-solving system is generated. The system is written in the PROLOG language and acts as a robot plan generator that is integrated with a path finding procedure in order to avoid collisions whilst plans are being implemented.

This paper, as an extension of an earlier paper, presents a geometrical representation and theoretical foundation of robot grasping that is affected by friction and by the magnitudes of normal contact forces. Grasp analysis and synthesis are based on the concepts of constraint cone (region), restraint cone and freedom cone in the force screw-space and location screw-space. In fact, the freedom cone is the aggregate of all the screws repelling and reciprocal to the grasp geometry (denoted by the restraint cone). Moreover, the constraint cone and the constraint region describe the effect of friction and the influence of the amplitudes of normal forces. The conditions of equilibrium grasp, stable grasp, form closure, force closure, and relative form closure are derived from the geometrical representation, and approaches are suggested for the design of adaptable fixtures, and for automatic grasp planning. It is shown that relative form closure is of great significance for frictional grasp, since it is possible to generate enough internal forces in the subspace spanned by the contacts and consequently produce corresponding friction forces in the reciprocal subspace.

The problem of finding an allowable object trajectory for a cooperating two-robot system is investigated. The purpose is to move an object from one point to another by firmly grasping it at two different points using two robotic hands. The major difficulty is caused by the fact that, unless the robots have true six degrees of freedom, the trajectories the object can follow are severely limited and, in general, are difficult to find. The method proposed in the paper is based on reformulating the problem as a nonlinear optimization problem with equality constraints in terms of the joint variables. The optimization problem is then solved numerically on a computer. The solution automatically gives the corresponding joint variable trajectories as well, thus eliminating the need for solving the inverse kinematic problem. The method has been successfully applied to a real experimental system.

Variable geometry truss manipulators (VGTM) are static trusses where the lengths of some members can be varied, allowing one to control the position of the free end relative to the fixed one. This paper deals with a planar VGTM consisting of a n–bay triangle-triangle truss with one variable length link (i.e. one DOF) per bay. Closed-form solutions to the forward, inverse, and velocity kinematics of a 3-DOF version of this VGTM are presented, while the forward and inverse kinematics of an n–DOF (redundant) one are solved by a recursive and an iterative method, respectively. A numerical example is presented.

In this paper a method of minimal neighborhood for cost optimal trajectory planning along prescribed paths is introduced. The method exploits the phase-plane approach. In the phase-plane, in an iterative procedure, subareas of search are built, called neighborings, which surround the current-best trajectory. In each iteration, in order to find the next-best trajectory, the dynamic programming (pruned to the subarea) is used. The method of minimal neighborhood makes the neighborings as small as possible and therefore speeds up computations maximally. The tests carried out on a model of the IRb-6 ASEA robot have shown that the method of minimal neighborhood is much faster than dynamic programming applied to the whole phase-plane, while preserving the quality of the resulting trajectory.

This paper presents current trends in robot programming. The open problems with current robot programming systems are outlined and indications for solutions are given. Since computer controlled robots have been introduced, the methodology of robot programming has seen a great deal of development. Two completely different approaches to robot programming have been considered in the past. On the one hand within the Artificial Intelligence community a lot of research has been done to provide robots with autonomous reasoning capabilities. On the other hand, the need to control industrial robots has pushed the development of simple but effective methods for robot programming. To put it simply, Artificial Intelligence researchers have taken a top-down approach trying to solve the difficult problem of reasoning and have assumed that all the rest was easy. Others have taken a bottom-up approach first trying to control robots and only later trying to incorporate intelligence. The complexity of industrial automation tasks requires programming systems more sophisticated that those in use today. Artificial Intelligence is the best candidate to create the next generation of robot programming systems.

A local-remote telerobot control system is described which is being developed for time-delayed groundremote control of space telerobotic systems. The system includes a local site operator interface for interactive command building and sequencing for supervised autonomy and a remote site: the Modular Telerobot Task Execution System (MOTES), to provide the remote site task execution capability. The local site system also provides stereo graphics overlay on video with interactive update of the remote environmental model. The operator selects objects in the environment to interact with and skill types to specify the tasks to be performed, such as grasping a module or opening a door.

The goal of a sophisticated robot programming system is the ability to accept natural language input, such as English, and be able to convert this to a series of robot commands that will achieve the desired robot movement. This paper first examines what is meant by the concept of natural language understanding. Then a bottom-up development of a model of a robot system identifies the levels of control in a robot system and the role of software at each of these. This is related to the various models of software hierarchy that have been proposed by different authors and an essential deviation from these at the object level (or sub-task level) is identified. Finally, a new route is proposed that should be taken in defining the levels of software if intelligent programming of robots is to be achieved.

This paper presents a new approach for simplifying dynamic equations of motion of robot manipulators by using a nondimensionalization scheme. With this approach the dynamic analysis is done in a nondimensional space. That is, it is required to establish a dimensionless coordinate system in which the dynamic equations of motion of manipulators are formulated. The characteristic parameters of the manipulators are then defined by choosing proper physical quantities as basic units for nondimensionalization. Within the nondimensional space the Lagrange method is applied to the manipulator to obtain a set of general dimensionless equations of motion. This dimensionless dynamic formulation of manipulators leads to an easier way to simplify the dynamic formulation by neglecting insignificant terms using the order of magnitude comparison. The dimensionless dynamic model and its simplified version of PUMA 560 robot are implemented using the proposed approach. It is found that the simplified dynamic model greatly reduces the computation burden of the inverse dynamics. Simulation results also show that the simplified model is extremely accurate. This implies that the proposed nondimensional simplification emethod is reliable.

A new method of optimal cost trajectory planning based on a graph searching algorithm is presented. Various heuristic functions which play the key role in the construction of effective algorithms for solving such problems are proposed. Some numerical examples are given based on the kinematic and dynamic models of a IRb-6 ASEA robot. This method rooted in AI has less computational complexity than a dynamical programming method applying to the same optimal-cost trajectory planning problem.