A master unit includes a welding DB in which prescribed motion data associated with an object to be welded is stored, a state sensor which measures, as welding state data, a situation of welding by a robot which executes welding in a real environment according to the prescribed motion data, a simulated environment which imitates the real environment and notifies a worker of the welding state data, and a motion control unit which receives, as an input, worker motion data indicating a motion of welding by the worker from the simulated environment, operates the robot in the real environment by using the worker motion data instead of the prescribed motion data, and records, as new prescribed motion data, the input worker motion data in the welding DB.

A training data screening device includes a data evaluation model, a data evaluator, a memory, and a training data screener. The data evaluation model is constructed by machine learning on at least a part of the collected data, or by machine learning on data different from the collected data. The data evaluator evaluates the input collected data using the data evaluation model. The memory stores the evaluated data, which is the collected data evaluated by the data evaluator. The training data screener screens the training data tier constructing the learning model from the evaluated data stored by the memory by an instruction of an operator to whom an evaluation result of the data evaluator is presented, or automatically screens the training data based on the evaluation result.

The present invention features a computer-implemented method of planning a processing path relative to a three-dimensional workpiece for a plasma arc cutting system coupled to a robotic arm. The method includes receiving input data from a user comprising (i) Computer-Aided Design (CAD) data for specifying a desired part to be processed from the three-dimensional workpiece, and (ii) one or more desired parameters for operating the plasma arc cutting system. A plurality of features of the desired part to be formed on the three-dimensional workpiece are identified based on the CAD data. The method also includes dynamically filtering a library of cut charts based on the plurality of features and the desired operating parameters to determine a recommended cut chart for processing the plurality of features. The method further includes generating the processing path based on the recommended cut chart and the plurality of features to be formed.

Proposed is a capturing apparatus. The capturing apparatus may include a setting unit configured to set environment information of a robot equipped with a camera; and a pattern unit configured to set a capturing pattern of the robot based on the environment information.

A robot control device includes an obtaining unit that obtains, from an image sensor that captures a workpiece group to be handled by a robot, a captured image, a simulation unit that simulates operation of the robot, and a control unit that performs control such that the captured image is obtained if, in the simulation, the robot is retracted from an image capture forbidden space, in which an image is potentially captured with the workpiece group and the robot overlapping each other, and which is set based on either or both a first space being the visual field range of the image sensor, and a columnar second space obtained by taking a workpiece region including the workpiece group or each of divided regions into which the workpiece region is divided, as a bottom area, and extending the bottom area to the position of the image sensor.

A differentiable simulator for simulating the cutting of soft materials by a cutting instrument is provided. In accordance with one aspect of the disclosure, a method for simulating a cutting operation includes: receiving a mesh for an object, modifying the mesh to add virtual nodes associated with a predefined cutting plane, optimizing a set of parameters associated with a simulator based on ground-truth data, and running a simulation via the simulator to generate outputs that include trajectories associated with a cutting instrument. Optimizing the set of parameters can include performing inference based on a set of ground-truth trajectories captured using sensors to measure real-world cutting operations. The inference techniques can employ stochastic gradient descent, stochastic gradient Langevin dynamics, or a Bayesian approach. In an embodiment, the simulator can be utilized to generate control signals for a robot based on the simulated trajectories.

The invention relates to a production planning method using a plurality of manufacturing devices (INTMA) according to which tasks (TD) of a work plan (BOP) are compared (MA) with manufacturing capabilities (SD) of the manufacturing devices (INTMA) and, depending on the one or more results (MAQ) of said comparison (MA), at least one or more manufacturing devices (INTMA) are commissioned to match their manufacturing capabilities (SD) with the task(s) (TD).

A dual-arm robot assembling system including a controlling unit, a GUI, a first robotic-arm, and a second robotic-arm is disclosed. The GUI provides a graphic program editing page, which provides multiple instruction blocks used for editing a graphical program executed by the assembling system. At least one of the first robotic arm and the second robotic arm is disposed with a point-teaching tool thereon. Before the controlling unit controls the two robotic arms to perform an assembling operation based on the graphical program, a manager may directly drag the two robotic arms through the point-teaching tool, so as to implement a point-teaching procedure for the two robotic arms. Therefore, the assembling system may accomplish the assembling operation through the two robotic arms with cooperative movement.

An example method for providing a graphical user interface (GUI) of a computing device includes receiving an input indicating a target pose of the robot, providing for display on the GUI of the computing device a transparent representation of the robot as a preview of the target pose in combination with the textured model of the robot indicating the current state of the robot, generating a boundary illustration on the GUI representative of a limit of a range of motion of the robot, based on the target pose extending the robot beyond the boundary illustration, modifying characteristics of the transparent representation of the robot and of the boundary illustration on the GUI to inform of an invalid pose, and based on the target pose being a valid pose, sending instructions to the robot causing the robot to perform the target pose.

A robot device, a method for training a robot control model, and a method for controlling a robot device. The method for training includes: supplying an image showing object(s), to a first and second prediction model to produce a first and second pickup prediction that has, for each pixel of the image, a first and second pickup robot configuration vector with an assigned first and second success probability; supplying the first and second pickup prediction to a blending model of the robot control model to produce a third pickup prediction that has, for each pixel of the image: a third pickup robot configuration vector that is a weighted combination of the first and second pickup robot configuration vector, and a third success probability that is a weighted combination of the first and second success probability; and training the robot control model by adapting the first and second weighting factors.