Solutions for multi-robot configurations are co-optimized, to at least some degree, across a set of non-homogenous parameters based on a given set of tasks to be performed by robots in a multi-robot operational environment. Non-homogenous parameters may include two or more of: the respective base position and orientation of the robots, an allocation of tasks to respective robots, respective target sequences and/or trajectories for the robots. Such may be executed pre-runtime. Output may include for each robot: workcell layout, an ordered list or vector of targets, optionally dwell time durations at respective targets, and paths or trajectories between each pair of consecutive targets. Output may provide a complete, executable, solution to the problem, which in the absence of variability in timing, can be used to control the robots without any modification. A genetic algorithm, e.g., Differential Evolution, may optionally be used in generating a population of candidate solutions.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using equipment-specific sample generators to automatically adapt a skill for execution in an operating environment. One of the methods includes obtaining a skill to be executed in an operating environment having one or more robots, wherein the skill defines a sequence of subtasks to be performed by the one or more robots in the working environment. A planning process is performed to generate a motion plan for the one or more robots, including obtaining, from an equipment-specific sample generator, a plurality of equipment-specific samples for a subtask of the skill, generating, from the plurality of equipment-specific samples, a plurality of candidate motion plans, and selecting, from the plurality of candidate motion plans, a motion plan for the one or more robots to perform the skill in the operating environment.

An off-line programming apparatus includes a model creation unit that creates three-dimensional models of a robot and a load, a storage unit that stores a dynamic parameter of the load, a graphic creation unit that creates a three-dimensional graphic representing the dynamic parameter based on the dynamic parameter, and a display unit that displays the three-dimensional models of the robot and the load and the three-dimensional graphic. The dynamic parameter includes inertia around three axes that are orthogonal to one another at a centroid of the load. The three-dimensional graphic is a solid defined by dimensions in three directions orthogonal to one another. The graphic creation unit sets a ratio of the dimensions in the three directions of the three-dimensional graphic to a ratio corresponding to a ratio of the inertia around the three axes.

The present invention relates to a collision-free path generating method for a robot and an end effector quipped thereon to move. The method includes steps of configuring a virtual working environment, containing a plurality of virtual objects at least including the robot, the end effector and a target object consisting of a plurality of basic members and mapped from a working environment in a reality, in a robot simulator; selecting a level of detail and a pre-determined shape for a collider covering the plurality of virtual objects to determine boundaries for the plurality of objects; randomly sampling a combination of robot configurations; and based on the determine boundaries and the randomly sampled combination of robot configurations, performing a heuristic based pathfinding algorithm to compute a collision-free path for the robot and the end effector quipped thereon to move to the target object accordingly.

A machine learning device that learns a control command for a machine by machine learning, including a machine learning unit that performs the machine learning to output the control command; a simulator that performs a simulation of a work operation of the machine based on the control command; and a first determination unit that determines the control command based on an execution result of the simulation by the simulator.

A system for parameter tuning for robotic manipulators is provided. The system includes an interface configured to receive a task specification, a plurality of physical parameters, and a plurality of control parameters, wherein the interface is configured to communicate with a real-world robot via a robot controller. The system further includes a memory to store computer-executable programs including a robot simulation module, a robot controller, and an auto-tuning module a processor, in connection with the memory. In this case, the processor is configured to acquire, in communication with the real-world robot, state values of the real-world robot, state values of the robot simulation module, simultaneously update, by use of a predetermined optimization algorithm with the auto-tuning module, an estimate of one or more of the physical, and said control parameters, and store the updated parameters.

Provided is a machine learning data generation device including: a virtual sensor input generator configured to generate a virtual sensor input, which is obtained by virtually generating a sensor input obtained as a result of performing sensing, by a sensor of a work machine, on a plurality of randomly piled subjects to be subjected to physical work by an operating machine of the work machine; a virtual operation command generator configured to generate a virtual operation command, which is obtained by virtually generating an operation command for the operating machine of the work machine; a virtual operation outcome evaluator configured to evaluate an outcome of the physical work in response to the virtual operation command in a virtual space; and a machine learning data generator configured to generate machine learning data based on the virtual sensor input, the virtual operation command, and the evaluation of the virtual operation outcome evaluator.

Systems and a method determine a sequence of joint values of an external axis along a sequence of targets. Inputs are received, including robot representation, tool representation, sequence of targets, kinematics of the axis joints, and/or type of robot-axis motion. For each target, it is generated at least one weight factor table representing, for each available configuration of the axis joint motion, a combined effort of the robot motion and the axis motion depending on the type of combined robot-axis motion. Valid weight factor values of the table are determined by simulating collision free trajectories for reaching the target. The sequence of joint values of the at least one external axis is determined by finding from the weight factor table a sequence of joint values for which the sum of their corresponding weight factors for reaching the target location sequence is minimized.

A system and method for automatically moving one or more items between a structure at a source location and a destination using a robot is provided. The system includes first and second vision systems to identify an item and to determine the precise location and orientation of the item at the source location and the precise location and orientation of the destination, which may or may not be in a fixed location. A controller plans the best path for the robot to follow in moving the item between the source location and the destination. An end effector on the robot picks the item from the source location, holds it as the robot moves, and places the item at the destination. The system may also check the item for quality by one or both of the vision systems. An example of loading and unloading baskets from a machine is provided.

An information processing apparatus includes a processor configured to simulate a virtual robot and a virtual image pickup apparatus moving in an interlocked manner with each other in a virtual space. The processor is configured to associate setting information of the virtual image pickup apparatus with a teaching point of the virtual robot.