Abstract: Methods and apparatus for complex assembly via autonomous robots using reinforcement learning action primitives are disclosed. An example apparatus includes a construction manager and a movement manager. The construction manager is to determine sequences of reinforcement learning (RL) action primitives based on object location goals and associated assembly goals determined for respective ones of objects depicted in an imaged assembly of objects. The movement manager is to command a robot to construct a physical assembly of objects based on the sequences of RL action primitives. The physical assembly of objects is to correspond to the imaged assembly of objects.

Abstract: Methods and systems are provided for positioning a remote sensor within a target object. An articulated robotic system is coupled to the remote sensor. A positioning system determines a position of the target object to be inspected and determines a first position of the remote sensor. A control system calibrates a virtual representation of the target object with respect to the position of the target object, and tracks movement of the remote sensor relative to the target object.

Abstract: A robot control apparatus includes a transformation-matrix generating unit that generates, on the basis of N first command values for learning per one robot, which are position-corrected command values for positioning a master robot and a slave robot on tracks of their respective robots during the synchronous driving, concerning the respective N command values for learning, transformation matrices among the command values for learning, a command-value storing unit that outputs a first command value for driving at each of M operation periods for defining the track of the master robot, a transformation-matrix-function generating unit and a command-value generating unit that interpolate the N transformation matrices and generate a transformation matrix at the each first command value for driving, and a command-value generating unit that causes the transformation matrices to act on the M first command values for driving and generates M second command values for driving of the slave robot.

Abstract: Apparatus and methods for training and controlling of, for instance, robotic devices. In one implementation, a robot may be trained by a user using supervised learning. The user may be unable to control all degrees of freedom of the robot simultaneously. The user may interface to the robot via a control apparatus configured to select and operate a subset of the robot's complement of actuators. The robot may comprise an adaptive controller comprising a neuron network. The adaptive controller may be configured to generate actuator control commands based on the user input and output of the learning process. Training of the adaptive controller may comprise partial set training. The user may train the adaptive controller to operate first actuator subset. Subsequent to learning to operate the first subset, the adaptive controller may be trained to operate another subset of degrees of freedom based on user input via the control apparatus.

Abstract: Robotic devices may be trained by a user guiding the robot along a target trajectory using a correction signal. A robotic device may comprise an adaptive controller configured to generate control commands based on one or more of the trainer input, sensory input, and/or performance measure. Training may comprise a plurality of trials. During an initial portion of a trial, the trainer may observe robot's operation and refrain from providing the training input to the robot. Upon observing a discrepancy between the target behavior and the actual behavior during the initial trial portion, the trainer may provide a teaching input (e.g., a correction signal) configured to affect robot's trajectory during subsequent trials. Upon completing a sufficient number of trials, the robot may be capable of navigating the trajectory in absence of the training input.

Abstract: An image acquisition unit acquires an image including an object, and a controller starts a visual servo using the acquired image, on the basis of at least one of an error in calibration, an error in installation of a robot, an error resulting from the rigidity of the robot, an error of a position where the robot has gripped the object, an error regarding imaging, and an error regarding a work environment. Additionally, the controller starts the visual servo when the distance between one point of a working unit of the robot and the object is equal to or greater than 2 mm.

Abstract: Apparatus and methods for arbitration of control signals for robotic devices. A robotic device may comprise an adaptive controller comprising a plurality of predictors configured to provide multiple predicted control signals based on one or more of the teaching input, sensory input, and/or performance. The predicted control signals may be configured to cause two or more actions that may be in conflict with one another and/or utilize a shared resource. An arbitrator may be employed to select one of the actions. The selection process may utilize a WTA, reinforcement, and/or supervisory mechanisms in order to inhibit one or more predicted signals. The arbitrator output may comprise target state information that may be provided to the predictor block. Prior to arbitration, the predicted control signals may be combined with inputs provided by an external control entity in order to reduce learning time.

Abstract: Apparatus and methods for arbitration of control signals for robotic devices. A robotic device may comprise an adaptive controller comprising a plurality of predictors configured to provide multiple predicted control signals based on one or more of the teaching input, sensory input, and/or performance. The predicted control signals may be configured to cause two or more actions that may be in conflict with one another and/or utilize a shared resource. An arbitrator may be employed to select one of the actions. The selection process may utilize a WTA, reinforcement, and/or supervisory mechanisms in order to inhibit one or more predicted signals. The arbitrator output may comprise target state information that may be provided to the predictor block. Prior to arbitration, the predicted control signals may be combined with inputs provided by an external control entity in order to reduce learning time.

Abstract: The present invention concerns an imitation learning method for a multi-axis manipulator (7,7?). This method comprises the steps of capturing, at a set of successive waypoints (10,11) in a teach-in trajectory (4) of a user-operated training tool, spatial data comprising position and orientation of the training tool (3) in a Cartesian space; selecting, from among said set of successive waypoints (10,11), a subset of waypoints (11) starting from a first waypoint (11) of said set of successive waypoints (10,11), wherein for each subsequent waypoint (11) to be selected a difference in position and/or orientation with respect to a last previously selected waypoint (11) exceeds a predetermined threshold; fitting a set trajectory (4?) in said Cartesian space to said selected subset of waypoints (11); and converting said set trajectory into motion commands in a joint space of said multi-axis manipulator (7,7?).

Abstract: Adaptive controller apparatus of a robot may be implemented. The controller may be operated in accordance with a reinforcement learning process. A trainer may observe movements of the robot and provide reinforcement signals to the controller via a remote clicker. The reinforcement may comprise one or more degrees of positive and/or negative reinforcement. Based on the reinforcement signal, the controller may adjust instantaneous cost and to modify controller implementation accordingly. Training via reinforcement combined with particular cost evaluations may enable the robot to move more like an animal.

Abstract: Robots may manipulate objects based on sensor input about the objects and/or the environment in conjunction with data structures representing primitive tasks and, in some embodiments, objects and/or locations associated therewith. The data structures may be created by instantiating respective prototypes during training by a human trainer.

Abstract: A control system and method generate torque comments for motion of a target system in response to observations of a source system. Constraints and balance control may be provided for more accurate representation of the motion as replicated by the target system.

Abstract: A robot arm includes a grip part which is structured to be separated from an end effector attached to the robot arm. When the grip part is gripped by the user and shifted, the robot arm shifts tracking the grip part. Further, the grip part includes contact sensors, and a tracking control method is switched according to the value of the contact sensors.

Abstract: A prediction system includes an acquisition device, an execution device, and a prediction device. The acquisition device is configured to acquire a value of electric resistance of a cable secured by fixing members to a pair of link members configured to rotate relative to each other through a joint. The execution device is configured to cause the joint to execute a predetermined operation. The prediction device is configured to predict a service life of the cable based on a change in the value of electric resistance acquired by the acquisition device while the execution device is causing the joint to execute the predetermined operation.

Abstract: A robot according to the present disclosure includes a support structure and at least three main arms mounted to be movable relative to the support structure, wherein the outer ends facing away from the support structure are movable to different spatial positions relative to the support structure and relative to each other. The robot further comprises connecting elements having the same lengths by means of which the outer end of each main arm can be connected at a defined distance from the outer end of the two adjacent main arms. The present disclosure also provides a method of calibrating a robot.

Abstract: A vacuum processing apparatus includes a robot connected to a vacuum container to carry a wafer on one of its two arms to or from a processing chamber; a unit to detect an amount of deviation of the wafer from a predetermined wafer mounting position on the arm that may occur when the robot carries the wafer into or out of the processing chamber; and an adjusting device to adjust the operation of the robot based on the detected amount of deviation. The adjusting device adjusts the robot operation based on the result of a teaching operation performed in advance. After being subjected to the initial teaching operation, the robot again undergoes a second teaching operation according to the information on the amount of wafer position deviation that is detected by moving the wafer in a predetermined transfer pattern, before the wafer processing is performed.

Abstract: A weld bead shaping apparatus including: a gouging torch for gouging an object to be shaped; a shape sensor for measuring a shape of the object; a slider apparatus and an articulated robot for driving the gouging torch and shape sensor; an image processing apparatus; and a robot controlling apparatus. The image processing apparatus includes: a shape data extracting unit extracting shape data of the object, from a measurement result obtained by the shape sensor; and a weld reinforcement shape extracting/removal depth calculating unit calculating a weld reinforcement shape of the weld bead from a difference between the shape data and a preset designated shape of the object, and calculating a removal depth by which gouging is performed, based on the weld reinforcement shape. The robot controlling apparatus controls the slider apparatus, the articulated robot, and the gouging torch based on the weld reinforcement shape and the removal depth.

Abstract: It is possible to perform robot motor learning in a quick and stable manner using a reinforcement learning apparatus including: a first-type environment parameter obtaining unit that obtains a value of one or more first-type environment parameters; a control parameter value calculation unit that calculates a value of one or more control parameters maximizing a reward by using the value of the one or more first-type environment parameters; a control parameter value output unit that outputs the value of the one or more control parameters to the control object; a second-type environment parameter obtaining unit that obtains a value of one or more second-type environment parameters; a virtual external force calculation unit that calculates the virtual external force by using the value of the one or more second-type environment parameters; and a virtual external force output unit that outputs the virtual external force to the control object.

Abstract: In order for a control system that interlocks with a teaching data input unit and a robot to control the robot, when teaching data including layer selection information that is input through the teaching data input unit is received, the control system generates a joint space path of a layer of any one of teaching data of a first layer and teaching data of a second layer according to the layer selection information of the teaching data. The control system controls the robot based on the generated joint space path.

Abstract: A robot system includes a robot, a control device, and a projection device. The control device is configured to receive area information on an area defining an operation of the robot. The projection device is configured to project the area onto an object adjacent to the robot based on the area information received by the control device.

Abstract: Adaptive controller apparatus of a robot may be implemented. The controller may be operated in accordance with a reinforcement learning process. A trainer may observe movements of the robot and provide reinforcement signals to the controller via a remote clicker. The reinforcement may comprise one or more degrees of positive and/or negative reinforcement. Based on the reinforcement signal, the controller may adjust instantaneous cost and to modify controller implementation accordingly. Training via reinforcement combined with particular cost evaluations may enable the robot to move more like an animal.

Abstract: In the control method for mobile parallel manipulators, kinematic singularity and redundancy are solved through joint limits avoidance and manipulability criteria. By taking the MPM self-motion into consideration due to its redundancy, the inverse kinematic is derived using a hybrid neuro-fuzzy system, such as NeFIK. The discrete augmented Lagrangian (AL) technique is used to solve the highly nonlinear constrained multi-objective optimal control problem. An adaptive neuro-fuzzy inference system (ANFIS)-based structure (based on the result of the AL solution) is used to solve the online trajectory planning of the MPM.

Abstract: In a control apparatus for a robot arm, the feedback rule generating section generates a feedback rule in accordance with relationship between a time point of generation of a stimulus and variation in behavior after elapse of a reaction time in taught data as detected by the information variation point detecting section and the behavior variation point detecting section. The motion generating section generates a motion of the robot arm based on motion information, an information variation point, a behavior variation point, and the feedback rule. The controller controls the motion of the robot arm.

Abstract: Teaching images are acquired at a plurality of separate teaching points on a running route extending from a running start position to a goal position, respectively, under a first light environmental condition and a light environmental condition different from the first light environmental condition, and the teaching images are stored. A present teaching image serving as a target for a robot body in a running direction at present is selected from the stored teaching images. A driving mechanism is controlled so as to increase the matching degree between the present teaching image and an actual image taken by a camera.

Abstract: The present invention relates to a method, such as a surgical method for assisting a surgeon for placing screws in the spine using a robot attached to a passive structure. The present invention also related to a method, such as a surgical method for assisting a surgeon for removing volumes in the body of a patient using a robot attached to a passive structure and to a device to carry out said methods. The present invention further concerns a device suitable to carry out the methods according to the present invention.

Abstract: In a teaching device for robot that moves a flexible body holding unit holding a flexible body and performs teaching of a fitting task of the flexible body to a thing to generate teaching data, a reaching decision unit decides that a front end of the flexible body reaches the thing, a curvature acquiring unit acquires information of evaluation of a degree of curvature of the flexible body when the reaching decision unit decides the reaching, and a notification unit performs predetermined notification about teaching on the basis of the acquired information of evaluation.

Abstract: An apparatus executes movement control that causes a robot arm equipped with a camera to move up to an object, thereby enabling a manipulator to move to an object quickly, accurately, and stably as a control system. Specifically, when the object is not detected, the apparatus executes teaching playback control to cause a manipulator to move along a path up to a target position set in advance based on a position of the object. When the object is detected, the apparatus defines a position closer to the object than the target position as a new target position, sets a new path up to the new target position, executes teaching playback control to cause the manipulator to move along the new path until a switching condition for switching the movement control is fulfilled. When the switching condition is fulfilled, the apparatus executes visual servo control.

Abstract: An off-line programming system (10) which includes a three-dimensional shape arranging unit (27) which fills in a curved surface or consecutive plurality of flat surfaces of a selected three-dimensional shape by selected operation patterns and arranges a three-dimensional shape in a virtual space so that the operation patterns will be projected on surfaces of the workpiece model, a working path preparing unit (28) which projects operation patterns on the surfaces of the workpiece model so as to prepare a working path of the tool, and a tool position/posture determining unit (29) which uses the prepared working path and normal direction of the surface of the workpiece model as the basis to automatically determined the position or position/posture of the tool model.

Abstract: Disclosed are a transport apparatus that holds and transports an object on a predetermined transport track using a transport portion provided at the leading end of an arm and is capable of acquiring the teaching information of a transport position using a normal transport operation, a position teaching method, and a sensor jig. A transmissive sensor (32) is provided in a sensor jig (30) such that the projection segments of an optical axis (41) and an optical axis (42) on a projection plane intersect with each other and neither the project segment of the optical axis (41) nor the projection segment of the optical axis (42) is aligned with the X-direction and the Y-direction. During a position teaching operation, the sensor jig (30) is provided so as to be held by a wafer transport portion (24), thereby detecting target members (51, 52).

Abstract: An apparatus for teaching a gripping device is disclosed. The gripping device includes an arm. The arm has a free end that can be freely moved in an operating region and includes a gripper. The gripper includes at least one, preferably two pairs of spaced apart gripping tongs consisting of an electrically conductive material. A control unit is included electrically connected to the tongs or a measuring body. A detecting means detects current in a control line. The apparatus uses a reference component having three reference surfaces to determine a reference coordinate system. By pre-specifying several points on the reference component that are to be approached, an automated determination of the position and alignment of a chuck or pallet can be utilized. A highly accurate analysis of a current signal is used by the apparatus for teaching the gripping device.

Abstract: A brain machine interface for control of prosthetic devices is provided. In its control, the interface utilizes parallel control of a continuous decoder and a discrete action state decoder. In the discrete decoding, we not only learn states affiliated with the task, but also states related to the velocity of the prosthetic device and the engagement of the user. Moreover, we not only learn the distributions of the neural signals in these states, but we also learn the interactions/transitions between the states, which is crucial to enabling a relatively higher level of performance of the prosthetic device. Embodiments according to this parallel control system enable us to reliably decode not just task-related states, but any “discrete action state,” in parallel with a neural prosthetic “continuous decoder,” to achieve new state-of-the-art levels of performance in brain-machine interfaces.

Abstract: A method for controlling a redundant robot arm includes the steps of selecting an application for performing a robotic process on a workpiece with the arm and defining at least one constraint on motion of the arm. Then an instruction set is generated based upon the selected application representing a path for a robot tool attached to the arm by operating the arm in one of a teaching mode and a programmed mode to perform the robotic process on the workpiece and movement of the arm is controlled during the robotic process. A constraint algorithm is generated to maintain a predetermined point on the arm to at least one of be on, be near and avoid a specified constraint in a robot envelope during movement of the arm, and a singularity algorithm is generated to avoid a singularity encountered during the movement of the arm.

Abstract: A system for understanding and storing knowledge and plans in a speakable, formal class of languages, allowing querying and retrieving of that knowledge and plan (conversation), and of deciding and acting (volition) upon such plans is provided. The system further allows the users to input data, instructions, questions, queries, imperatives, and plans without structure or rules over and above the actual grammar of the class of languages. The system responds in words, phrases, complete sentences, or actions, as applicable, and its output is bound by the rules of logic, rather than correlation or likely meaning.

Abstract: A computer program product and an apparatus for preparing a moving program for controlling the operation of a working robot which can move a known working apparatus relative to a workpiece and which can perform desired work on the workpiece. Movement information of the working apparatus may be input to a text entry screen on a character basis. Movement information of the working apparatus may also be input via a figure entry screen as a path on a two-dimensional plane in correlation with height information. The movement information that is input on the text entry screen is output in real time as the path on the two-dimensional plane and the height information thereof on the figure entry screen. The movement information that is input on the figure entry screen is output in real time to the text entry screen on the character basis.

Abstract: Disclosed is a teaching and playback method using a redundancy resolution parameter determined in conjunction with a joint structure, for a robot, and a method to apply analytic inverse kinematics to a robot having an elbow with an offset and a computer-readable medium of controlling the same. A reference plane variable with the joint structure is generated and an angle between the reference plane and an arm plane of the robot is used as the redundancy resolution parameter. The robot is taught and its operation is played back in differential inverse kinematics or analytic inverse kinematics using the resolution redundancy parameter.

Abstract: A positioning apparatus and a working system which can position a workpiece with better accuracy at a lower cost and in a reduced space even when the workpiece is heavy are provided. The positioning apparatus has 3 robot arms (1)-(3) each having an arm and an actuator which drives the arm, a mounting platform (5) which is supported by the ends of the three robot arms (1)-(3), a securing jig (6) which secures a workpiece W which is mounted on the mounting platform (5) to the mounting platform (5), and a controller (9) which controls the operation of each of the actuators.

Abstract: There is provided a method of post-correction of a 3D feature point-based direct teaching trajectory, which improves direct teaching performance by extracting shape-based feature points based on curvature and velocity and improving a direct teaching trajectory correction algorithm using the shape-based feature points. Particularly, there is provided a method of post-correction of a 3D feature point-based direct teaching trajectory, which makes it possible to extract and post-correct a 3D (i.e., spatial) trajectory, as well as a 2D (i.e., planar) trajectory, with higher accuracy.

Abstract: A robotic system includes a robot adapted for moving a payload in proportional response to an input force from an operator, sensors adapted for measuring a predetermined set of operator input values, including the input force, and a controller. The controller determines a changing stiffness value of the operator using set of operator input values, and automatically adjusts a level of control sensitivity over the robot using the stiffness value. The input values include the input force, a muscle activation level of the operator, and a position of the operator. A method of controlling the robot includes measuring the operator input values using the plurality of sensors, processing the input values using the controller to thereby calculate the stiffness value, and automatically adjusting the level of control sensitivity over the robot using the stiffness value. A specific operator may be identified, with control sensitivity being adjusted based on the identity.

Abstract: Certain embodiments of the present invention are generally directed to devices and methods for using pressure measurements to teach a robot a teaching point location. In certain embodiments, a method includes measuring pressure at multiple points across a target. A location of a robot teaching point is determined based on the measured pressure.

Abstract: A robot system according to one aspect of an embodiment includes a robot and an instructing module. The robot holds one of a plurality of feed materials used for processing a workpiece. The instructing module gives instructions to the robot, when the feed materials are used for processing the single workpiece, for an operation in which the feed material held last in the previous round of processing a workpiece is used first in the subsequent round of processing a workpiece.

Abstract: The invention relates to a surgery assistance system for guiding a surgical instrument. The surgical instrument (3) is fastened lo an arm system (10, 12, 14), the tip (S) of the surgical instrument (3) can be moved in a controlled manner by means of the arm system (10, 12, 14) in a Cartesian patient coordinate system (PKS), one of the three spatial axes (x, y, z) of the Cartesian patient coordinate system (PKS) extends through the surgical opening or the trocar point (T) receiving the surgical instrument (3). Advantageously, the angle of inclination (w) of the surgical instrument (3) is determined with respect to the spatial axis (z) of the Cartesian patent coordinate system (PKS) extending through the trocar point (T), the angle of inclination (w) so determined is compared with a predetermined set angle of inclination (ws) for the purpose of guiding the tip (S).

Abstract: A robot system includes a robot control device connected to a plurality of robots each having a plurality of joints and a robot teaching device configured to communicate with the robot control device so as to teach and operate the robots. The robot control device automatically select one of the robots to be operated in accordance with a change in a display state of a display of the robot teaching device.

Abstract: Teaching images are acquired at a plurality of separate teaching points on a running route extending from a running start position to a goal position, respectively, under a first light environmental condition and a light environmental condition different from the first light environmental condition, and the teaching images are stored. A present teaching image serving as a target for a robot body in a running direction at present is selected from the stored teaching images. A driving mechanism is controlled so as to increase the matching degree between the present teaching image and an actual image taken by a camera.

Abstract: A vacuum processing apparatus includes a robot connected to a vacuum container to carry a wafer on one of its two arms to or from a processing chamber; a unit to detect an amount of deviation of the wafer from a predetermined wafer mounting position on the arm that may occur when the robot carries the wafer into or out of the processing chamber; and an adjusting device to adjust the operation of the robot based on the detected amount of deviation. The adjusting device adjusts the robot operation based on the result of a teaching operation performed in advance. After being subjected to the initial teaching operation, the robot again undergoes a second teaching operation according to the information on the amount of wafer position deviation that is detected by moving the wafer in a predetermined transfer pattern, before the wafer processing is performed.

Abstract: Provided is a robot capable of appropriately adjusting a position and the like of a main body in view of executing a specified task involving an interaction with a target object. While the position and posture of the main body (10) are being controlled according to a second target path, the robot (1) moves from a first specified area to a second specified area and stands there. In this state, a second position deviation (=the deviation of the position of the main body from a second target path) and a second posture deviation (=the deviation of the posture of the main body from a second target posture) are determined. According to the determination result, the second target path is corrected so that the subsequent position deviation and the like may be smaller.

Abstract: The safety of an operator which may be endangered by an erroneous instruction by the operator or a robot control system is ensured by making more stringent a condition regarding the separation of the operator from the vicinity of a robot when an operation program of the robot is activated. An interlock to which a condition regarding activation of the operation program of the robot is added is provided in a feeding unit which is connected to a robot controlling unit by wireless connection for charging a teaching unit, so as to provide a robot system which improves the safety of the operator.

Abstract: A method for adjusting a program including program instructions for controlling an industrial robot to carry out work at a plurality of target points on a work object. The robot includes a tool having two arms adapted to clamp the work object and at least one of the arms is arranged movable relative the other arm in an opening and a closing direction, a manipulator adapted to hold the tool or the work object, and a controller controlling the movements of the manipulator and the tool arm and configured to switch between a normal control mode and a compliant control mode in which the manipulator has a reduced stiffness in at least one direction. The method includes moving the manipulator and the tool according to the program instructions until one of the target points is reached.

Abstract: The parallel kinematic machine (PKM) trajectory planning method is operable via a data-driven neuro-fuzzy multistage-based system. Offline planning based on robot kinematic and dynamic models, including actuators, is performed to generate a large dataset of trajectories, covering most of the robot workspace and minimizing time and energy, while avoiding singularities and limits on joint angles, rates, accelerations and torques. The method implements an augmented Lagrangian solver on a decoupled form of the PKM dynamics in order to solve the resulting non-linear constrained optimal control problem. Using outcomes of the offline-planning, the data-driven neuro-fuzzy inference system is built to learn, capture to and optimize the desired dynamic behavior of the PKM. The optimized system is used to achieve near-optimal online planning with a reasonable time complexity. The effectiveness of the method is illustrated through a set of simulation experiments proving the technique on a 2-degrees of freedom planar PKM.

Abstract: It is possible to perform robot motor learning in a quick and stable manner using a reinforcement learning apparatus including: a first-type environment parameter obtaining unit that obtains a value of one or more first-type environment parameters; a control parameter value calculation unit that calculates a value of one or more control parameters maximizing a reward by using the value of the one or more first-type environment parameters; a control parameter value output unit that outputs the value of the one or more control parameters to the control object; a second-type environment parameter obtaining unit that obtains a value of one or more second-type environment parameters; a virtual external force calculation unit that calculates the virtual external force by using the value of the one or more second-type environment parameters; and a virtual external force output unit that outputs the virtual external force to the control object.

Abstract: A robot system includes a robot, and a robot controller configured to control operation of the robot. The robot includes two arms including tools at their tip ends, and a body that supports the two arms and allows them to swing. The robot controller controls the robot so as to keep the tool provided at the tip end of at least one arm of the two arms at the same position in the same posture during a time period between before and after a swing.