Abstract: A shared system for robot control including integration of the human and autonomous input modalities for an improved control.Autonomously planned motion trajectories are modified by a teleoperator to track unmodelled target motions, while nominal teleoperator motions are modified through compliance to accommodate geometric errors autonomously in the latter. A hierarchical shared system intelligently shares control over a remote robot between the autonomous and teleoperative portions of an overall control system. Architecture is hierarchical, and consists of two levels. The top level represents the task level, while the bottom, the execution level.In space applications, the performance of pure teleoperation systems depend significantly on the communication time delays between the local and the remote sites. Selection/mixing matrices are provided with entries which relfect how each input's signals modality is weighted.

Abstract: An artificial neural network, which has a plurality of neurons each receiving a plurality of inputs whose effect is determined by adjust able weights at synapses individually connecting the inputs to the neuron to provide a sum signal to a sigmoidal function generator determining the output of the neuron, undergoes memory modification by a steepest-descent method in which individual variations in the outputs of the neurons are successively generated by small perturbations imposed on the sum signals. As each variation is generated on the output of a neuron, an overall error of all the neuron outputs in relation to their desired values is measured and compared to this error prior to the perturbation. The difference in these errors, with adjustments which may be changed as the neuron outputs converge toward their desired values, is used to modify each weight of the neuron presently subjected to the perturbation.

Abstract: A computer system and process of operating a computer having a memory for storing knowledge which includes a reflection base and special knowledge. The reflection base contains complete instructions which can be executed by processors of known type and meta-knowledge about the instructions. The instructions provide components of new knowledge and the meta-knowledge is applied to the construction of compositions of the instructions. The special knowledge contains knowledge in a field of application. The special knowledge is changed and extended on the basis of the special knowledge itself wherein compositions of the instructions are constructed from the reflection base and the special knowledge, the execution of these compositions of instructions is controlled on the basis of the special knowledge, and components of the changed and extended special knowledge are generated from parts of the special knowledge and parts of the compositions of instructions.

Abstract: This is a recurrent or feedforward analog neural network processor having a multi-level neuron array and a synaptic matrix for storing weighted analog values of synaptic connection strengths which is characterized by temporarily changing one connection strength at a time to determine its effect on system output relative to the desired target. That connection strength is then adjusted based on the effect, whereby the processor is taught the correct response to training examples connection by connection.

Abstract: A force and moment which a manipulator receives from an external environment is detected, and this detected value is multiplied by a gain inversely proportional to a virtual spring constant set by a tool coordinate system, and the product is further converted to a value in each joint coordinate system of the manipulator so as to determine a detection torque. A command value of force and moment is converted to a value in each joint coordinate value of the manipulator in the same way as described above, so as to determine a command torque. A difference between a position command value and a position detection value is multiplied by a virtual spring constant in each joint coordinate system of the manipulator obtained by converting the virtual spring constant, and a differential torque is obtained by converting the aforementioned difference to a force and moment corresponding to the difference.

Abstract: An articulated robot such as an arc welding robot includes a plurality of articulations and an end effector having a tool center point. To teach the robot, a coordinate transformation matrix for transforming coordinates between the articulations is determined from condition data of the articulations. A present position of the tool center point is calculated based on the coordinate transformation matrix and a moving vector is determined from the present position of the tool center point to a next position thereof. Thereafter, a new coordinate transformation matrix for transforming coordinates between the articulations is determined so that the posture of the end effector at the next position of the tool center point, which is calculated based on the first-mentioned coordinate transformation matrix and the moving vector, will coincide with the posture of the end effector at the present position of the tool center point.

Abstract: After positional information of the master arm is transformed into first generalized world coordinates corresponding to positional information representing a calculated position of the slave arm, when a signal instructing an interruption of the associated operation between the master arm and the slave arm is received and then the positional relationships between the master arm and the slave arm connected thereto by use of an imaging device are updated, the first generalized world coordinates are transformed into the second generalized world coordinates based on the new positional relationships so as to restart the associated operation therebetween. The optimal control of the manipulator is conducted through matrix operations related to the change-over between imaging devices, the reset operation conducted in an area of the movable range of the master arm, and the alteration of the ratio between the movements of the associated operation.

Abstract: A knowledge inferential processing apparatus comprises an external interface section for inputting and outputting information between the knowledge inferential processing apparatus and the exterior thereof; a rule base memory for representing knowledge as a combination of a plurality of rules constructing a premise conditional section and a conclusive section as a pair, and storing the represented knowledge; a rule pointer table for storing a rule pointer group indicative of a sequence for applying the rules within the rule base memory; an inferential processing section for interpreting and executing the rules and judging whether or not the matching is formed with respect to the respective rules; a memory section for storing fact data sent to the inferential processing section and conclusive data generated during the inferential processing; an inferential hysteresis recording section for recording a rule pointer of the matched rule in the matched sequence with respect to the inferential processing results exe

Abstract: The station comprises a revolving table for receiving, positioning and clamping in an operating position a pallet of the type provided with adjustable anchoring elements for components of clamping fixtures, a magazine for modular components of clamping fixtures and a robot capable of executing adjustment functions of the anchoring elements of pallet, of the clamping and transfer of the modular components and of the application to and separation from the components themselves with respect to anchoring elements of pallet. There is provided a control centre with a computer, whereby just one operator may automatically and repetitively execute the work for the entire station.

Abstract: An automatic program synthesizer is disclosed.

Abstract: A self-extending shape neural-network is capable of a self-extending operation in accordance with the studying results. The self-extending shape neural-network has initially minimum number of the intermediate layers and the number of the nodes (units) within each layer by the self-extension of the network construction so as to shorten the studying time and the discriminating time. This studying may be effected efficiently by the studying being directed towards the focus when the studying is not focused.

Abstract: A peripheral control device applicable in robotics for control of educational robots operated by stepper motors performs reliably positioning tasks of a great number of steps with a great number of axes controlled with different velocity and realizes, by means of digital inputs and outputs, a reliable connection with the technological equipment with minimum participation of the microprocessor which controls it. The device is controlled by a microprocessor and contains a buffer for addresses, a second buffer for control signals, a bidirectional buffer for data, a decoder, a register for digital inputs, a second register for digital outputs, a third register for availabilities, a fourth register for directions, a fifth register for movements, and a sixth register for the control of a programmable counter. The programmable counter sets the rotation velocities of the stepper motors.

Abstract: A rotary trowel is provided for a working robot which can freely run and the concreted floor is flattened and finished by the rotary trowel while the robot runs on the concreted floor. The working robot is automatically operated in accordance with a zigzag running pattern which was set in a desired working range which is determined by a width and a depth. When the teaching mode is set, the first straight running and the first turning operation of a zigzag running pattern which will be set are executed by the manual operation. Thus, a straight running distance and turning direction which were detected by the manual operation are set as initial values and the zigzag running pattern is automatically set. Subsequently, the automatic operation is performed in accordance with the zigzag running pattern which was automatically set.

Abstract: A robot controller for having a workpiece held by a robot and having the workpiece machined by moving it relative to a stationary tool. The robot controller can control the velocity of relative motion, against the tool, of a machining point on the workpiece. At each teaching point, a transformation matrix is calculated to convert a position matrix of the flange center as teaching data (i.e., 4.times.4 matrix giving position and attitude in homogeneous coordinates) into a position matrix of the tool tip. Between two adjacent teaching points, interpolation points are set according to a specified velocity. Transformation matrices applicable to the two teaching points are put to interpolation calculations to provide a transformation matrix at each interpolation point. An inverse matrix of the transformation matrix at each interpolation point is applied to the position matrix for the tool tip. This allows the position matrix for the flange center keyed to each interpolation point to be calculated.

Abstract: The brightness and the contrast are optimized in a video image which is used for controlling a manipulator unit. For this purpose, the gain/offset control means of a video camera producing the video image, is controlled using an image control unit in a manner such that the image data of an object can be processed by the image control unit in optimum manner. The manipulator unit may comprise an industrial robot manipulating a workpiece or a tool of a machine tool acting upon a workpiece. The manipulator unit can also be constructed as a target tracking platform for tracking a target which is imaged by the video camera.

Abstract: A moving type working robot comprises a vehicle provided with a means for setting a traveling course between a position of stopping its working operation and an accumulating/replenishing place on the basis of positional information of the position where the working was stopped and the position of the place where the material loaded on the vehicle is to be accumulated/replenished. The positional information is stored in storing means and a control means is provided for traveling the vehicle from said accumulating/replenishing place to said position at which the working had been stopped, or vice versa, along said traveling course when the amount of the material loaded on the vehicle has reached a predetermined value.