At present, the procedure which is most widely used to identify an object consists of displacing its image in order that it be as close as possible to a reference pattern, and then to make a decision based upon the value of an Euclidean distance. This decision may be either in a deterministic framework or in a statistical one. The transformations so involved are translation and rotation. It is shown that, if one takes for granted the invariance of the amount of information involved by the object considered as an informational source, then there is another class of transformations, referred to as deformation, which should be also taken into account. The question is examined, and its consequences in the design of artificial vision are outlined in both the deterministic and stochastic cases.

The analysis of the gait mode is an important step for constructing a walking robot. In this study an optimal turning gait of a four-legged walking robot is proposed, which maximizes walking speed keeping walking stability larger than a certain required value. To obtain an optimal gait algorithm, for body trajectory of the robot, maximum walking speed, feasible gait modes and stability margin are analyzed for a quadruped walking robot designed in this study. The proposed gait is sufficiently flexible, as to permit crab walking, turning walking and a pure rotational walking about a robot's geometric center. The validity of the proposed gait algorithm is confirmed by many simulations.

Recent hardware advances for robot accessories include self changing grippers. A universal wrist has the capability of accessing and pneumatically attaching itself to a limited number of grippers that can be stored in a magazine. This paper addresses the determination of self changing gripper characteristics to permit the grasping of a wide variety of geometric shapes with a limited number of different gripper types.

We describe a robot capable of performing all procedures necessary to carry out a complete stereotactic neurosurgical operation under the control and supervision of a surgeon. The operation consists of the introduction of a small probe with diameter 2–3 mm through a hole without trepanation. The robot has been built and is now being tested and evaluated. The accompanying control software as well as various medical probes are either in development or partially tested. The installation will be able to carry out a complete intervention under the surveillance of a computed tomography scanner. In this article we emphasize the design choices required to eliminate gearing backlash in a crucial degree of freedom.

The payload variation has been one of the principal reasons in reducing path tracking accuracy and complicating controller design. In this paper, the trajectory and input sensitivities for the payload variation are investigated for three different robot configurations which include unbalanced, inaccurately balanced and balanced configurations. Based upon the sensitivity theory, simulation studies were made to evaluate the sensitivities of these configurations with respect to payload variations. The simulation results are discussed in detail and compared for the three configurations.

An optimization technique based on the well known Dynamic Programming Algorithm is applied to the motion control trajectories and path planning of multi-jointed fingers in dextrous hand designs. A three-fingered hand with each finger containing four degrees of freedom is considered for analysis. After generating the kinematics and dynamics equations of such a hand, optimum values of the joints torques and velocities are computed such that the finger-tips of the hand are moved through their prescribed trajectories with the least time or/and energy to reach the object being grasped. Finally, optimal as well as feasible solutions for the multi-jointed fingers are identified and the results are presented.

After elucidating the terms ‘mining’ and ‘robot’, a historical background to the problem of mining is described, with emphasis on the use of technology. Various reasons for the increasing utilisation of technology are discussed, as regards the mechanisation of existing operations and those unachievable without new technology. It is asserted that the mining industry has a number of particular features that make increasing mechanisation especially attractive. The point is made that the ultimate stage of the historical tendency towards the use of new technology is fully automatic mining, a goal that may not be realised, however, in the next decade.

The above historical discussion is followed by a critical appraisal of the conservative nature of the Australian industry and the study of the various factors that contribute to the slow uptake of new technology in mining.

Attention is then concentrated on the general problem of robots and mining. It is shown that worldwide there were no applications of robots to mining up to the end of 1981. A wide study has been carried out to identify some uses of robots in mining, but without success.

The next topic to be discussed is the widespread confusion about robots and teleoperators. The fundamental difference between the two is underlined, and particular attention is paid to Thring's teleoperator mining concept (telechiric mining). It is emphasized that telechiric mining will have no significant impact upon Australian mining in this decade.

These pessimistic asssessments are followed by claims that in some areas robotic concepts, though not robots themselves, could have a considerable effect on mining automation in the 1980's, both in Australia and elsewhere. The following projects are enumerated as possible candidates for applying robotic concepts: Surface mining; Mine development; Underground coal transport; Coal winning; Coal preparation. In order to make progress, it is recommended that a robotics expert ought to be included in the interdisciplinary teams studying the problem of mining automation.

The last part of this paper is devoted to Australia's need for robotic mining. A list of systems involving robotic concepts is presented that may be realised in the current decade. It is maintained that Australia ought to develop a few of such systems for the domestic and international markets, and that the short-term needs of the Australian mining industry are fundamentally different from those of other Australian industries as regards robots. Hence, decisions regarding robots for mining should be made independently from those appertaining to automation problems of other industries.

This work concerns the class of wheeled mobile robots with single axis front steering. Because of the relatively low speed of these vehicles their dynamic properties have little effect on their path tracking behaviour. Their motion is, moreover, on a flat environment and can be assumed two dimensional.

The kinematics of motion of such a vehicle can be utilized for design of control strategies for maintaining path following stability. In this paper, optimal control technique is implemented to such a system. First, the state space equations governing the motion are derived. The orientation error and the offset from a path form the states and the steering angle is the control input to the system. An optimal controller to minimize the two errors and the steering angle is then sought. The conditions for the existence of the feedback matrix are discussed. The controller structure is defined in terms of the forward speed of the vehicle and, thus, has the advantage of being flexible for speed changes. Numerical illustrative examples, however, demonstrate that variation of the speed has no effect on the controller structure.

A modeling method for robots is proposed, in which a convex polyhedron is represented as a set of inequalities and a robot is represented as a union of convex polyhedrons. With this method, collision between robots can be detected by solving a set of linear programming problems at every sampling instant. By detecting possible collision at every sampling instant, a directed graph for robots is created. The motion planning problem of robots is then transformed into a path searching problem in the directed graph and can be solved by exisiting searching algorithms.

The production system has been used extensively in Artificial Intelligence systems due, possibly, to a lack of a prior procedural orientation towards the knowledge embedded within it. As a result, this paradigm is especially useful in modeling domains in which a strong procedural correlation of data would not naturally appear. The production system is therefore an appropriate tool for designing models of systems of independent processes whose interactions can be defined in terms of system state.

We are interested in modeling continuous, concurrent processes for simulating robot activities, and present a description of a system which is capable of representing such processes as symbolic expressions within a production system database. This system implements a world model which acts as a continuous function of time, and a set of rules written in a language capable of specifying time-dependent properties of the model. In addition, rules may specify new rules, thus permitting processes to be mapped out over time as sequences of events.

The methodology presented in this paper is an attempt to utilize the power of symbolic programming in a design for world modeling, and a characterization of the requirements for applying such systems to problem domains which contain a continuous parameter, such as time. We believe that the technique of embedding time used here is appropriate for expressing the dynamic evolution of these models.