A grasping task is often the first stage to be carried out in an assembly or handling task. This paper points out one aspect of the grasping problem: How to manipulate an industrial part with robots in order to bring it to the right location and orientation according to a desired task (assembly, storage, movement,…)?

The method of parametric optimization is applied to the energy analysis of the motion of three-link manipulators. The quality criterion is the energy consumption along 1 m of the path. It consists of two parts: 1) The sum of the products of joint moments and relative joint velocities (an equivalent of the mechanical work); 2) a quadratic form of joint velocities (an analogue of dissipation losses). Typical spatial motions are studied: the hand of the robot moves along a straight line from an initial position to another one, and returns to the start. Several velocity function shapes (parabolic, sinusoidal, and triangular), both symmetric and non-symmetric ones, are considered. The dependence of the energy consumption on the velocity form, on the trajectory parameters and on mechanical and geometrical characteristics of links is discussed.

Much of the work of Artificial Intelligence concerns the representation and symbolic manipulation of various domains. One problem shared by existing systems is their limitation to discrete states and state changes, brought about by the linking of declarative states with points in time. With this identification, models of continuous processes must forego such declarative states, resulting in a loss of decomposability that limits the amount of complexity that can be handled by a given amount of computation.

In this paper we develop a model for robot problem solving over a domain of continuous actions. The proposed world model consists of a vector of piecewise defined functions of time, i.e. functions that can be represented using finite sets of expressions and finite partitions of time intervals. The critical points of these functions are represented symbolically as discrete events. The goal is represented as a point through which the world model vector must thread at some point in model time. World model functions are transformed by adding events at various discrete points in time which transform the trajectories in the direction of the goal. As events are added, formulas relating the event times are added to the description of the world. The result is a partial order of events which represents a plan for “forcing” the world function to achieve the goal.

The relationships between this model and the classical problem-solving systems (e.g. STRIPS) are utilized in order to produce a planning procedure. This procedure is applied to a robot problem which can be solved using the model. We conclude by presenting an analysis and prospectus of considerations for an actual implementation of these ideas.

A new elastic structure of a robot wrist, which exhibits a six decree of freedom flexural capability, is described. Force and compliance analysis is made, and algorithms for geometrical parameter synthesis are proposed. Owing to the technological simplicity of the designed structure, it is possible to construct cost-effective compliance devices. The intended use is to permit the assembly of close tolerance parts with industrial robots having relatively low positional accuracy. It could also be applied to other manual or automatic matching operations.

The author makes the point that progress is slow, though inexorable, and one can discern an upward drift of unemployment, even in good times. After analysing the various views concerning this phenomenon, the author points to two major problems, viz. the transitional one of the redundant heavy manual worker and the long-term one of employment for the unskilled. He then considers these problems from the point of view of automation and robotics with particular reference to the reaction of the human society to the qualitative changes in the manufacturing industries.