Abstract: Systems and methods for training a neural network to predict states of a robotics device are disclosed. Robotics data is received for a robotics device, including indications of a set of components, a digital simulation of the robotics device, and measurement data received from a sensor associated with the robotics device. The set of components includes an actuator and a structural element. A training dataset is generated using the received robotics data. Generating the training dataset includes comparing the measurement data with simulated measurement data based on the digital simulation. A neural network is trained using the generated training dataset to modify the digital simulation of the robotics device to predict a state of the robotics device, such as a position, motion, electrical quantity, or other. When trained, the neural network is applied to predict states of the robotics device or a different robotics device.

Abstract: A new controller for use in robots with kinematic loops as well as in most other types of robots (such as those with fully actuated kinematic trees). The controller includes an inverse kinematics (IK) module that implements a versatile IK formulation for retargeting of motions, including expressive motions, onto mechanical systems (i.e., robots with loops and/or without loops). Further, the controller is configured to support the precise control of the position and orientation of end effectors and the center of mass (CoM) (such as of walking robots). The formulation of the algorithms carried out by the IK module safeguards against a disassembly when IK targets are moved outside the workspace of the robot. A regularizer is included in the controller that smoothly circumvents kinematic singularities where velocities go to infinity.

Abstract: In one example, a robotic system is disclosed that includes a plurality of components coupled together, a plurality of motors operable to move the plurality of components, a controller in electrical communication with the plurality of motors to generate control signals to actuate movement of the plurality of components, wherein the controller is configured to: receive a first set of control signals operative to generate a defined motion for the plurality of components, analyze the first set of control signals to determine a second set of control signals operative to define a retargeted motion for the plurality of components, wherein the retargeted motion suppresses vibrations of the plurality of components as compared to the defined motion, and provide the second set of control signals to the plurality of motors to actuate the retargeted motion by the plurality of components.

Abstract: A robot control method, and associated robot controllers and robots operating with such methods and controllers, providing real-time vibration suppression. The control method involves learning to support real-time, vibration-suppressing control. The method uses state-of-the-art machine learning techniques in conjunction with a differentiable dynamics simulator to yield fast and accurate vibration suppression. Vibration suppression using offline simulation approaches that can be computationally expensive may be used to create training data for the controller, which may be provide by a variety of neural network configurations. In other cases, sensory feedback from sensors onboard the robot being controlled can be used to provide training data to account for wear of the robot's components.

Abstract: A robot control method, and associated robot controllers and robots operating with such methods and controllers, providing computational vibration suppression. Given a desired animation cycle for a robotic system or robot, the control method uses a dynamic simulation of the physical robot, which takes into account the flexible components of the robot, to predict if vibrations will be seen in the physical robot. If vibrations are predicted with the input animation cycle, the control method optimizes the set of motor trajectories to return a set of trajectories that are as close as possible to the artistic or original intent of the provider of the animation cycle, while minimizing unwanted vibration. The new control method or design tool suppresses unwanted vibrations and allows a robot designer to use lighter and/or softer (less stiff) and, therefore, less expensive systems in new robots.

Abstract: A new controller for use in robots with kinematic loops as well as in most other types of robots (such as those with fully actuated kinematic trees). The controller includes an inverse kinematics (IK) module that implements a versatile IK formulation for retargeting of motions, including expressive motions, onto mechanical systems (i.e., robots with loops and/or without loops). Further, the controller is configured to support the precise control of the position and orientation of end effectors and the center of mass (CoM) (such as of walking robots). The formulation of the algorithms carried out by the IK module safeguards against a disassembly when IK targets are moved outside the workspace of the robot. A regularizer is included in the controller that smoothly circumvents kinematic singularities where velocities go to infinity.

Abstract: A robot control method, and associated robot controllers and robots operating with such methods and controllers, providing real-time vibration suppression. The control method involves learning to support real-time, vibration-suppressing control. The method uses state-of-the-art machine learning techniques in conjunction with a differentiable dynamics simulator to yield fast and accurate vibration suppression. Vibration suppression using offline simulation approaches that can be computationally expensive may be used to create training data for the controller, which may be provide by a variety of neural network configurations. In other cases, sensory feedback from sensors onboard the robot being controlled can be used to provide training data to account for wear of the robot's components.

Abstract: A system providing dynamic balancing in a robotic system. The system includes memory storing a definition of a robot and storing an input animation for the robot specifying motion of components of the robot. A simulator performs a dynamic simulation of the robot performing the input animation including modeling a first set of the components as flexible components and a second set of the components as rigid components. Each of the flexible components is coupled at opposite ends to one of the rigid components. An optimizer generates a retargeted motion for the components to provide dynamic balancing of the robot performing the retargeted motion. The optimizer generates the retargeted motion by transforming forces acting on the robot to a local contact frame rigidly moving with the robot. The optimizer generates the retargeted motion so a zero-moment point of the robot lies in a support area of the robot's feet.

Abstract: A system for performing simulation-based material characterization includes a computing platform having a hardware processor and a system memory storing a software code. The hardware processor executes the software code to obtain a result of a physical test performed on a material, selects a parameterized model of the material based on the obtained result, and performs a simulation of the physical test using the parameterized model to generate a simulated result. The hardware processor further executes the software code to compare the simulated result with the obtained result of the physical test on the material, and adjusts one or more parameter value(s) of the parameterized model, based on the comparison, to improve the simulated result, and predict, after adjusting the parameter value(s), one or more characteristics of the material based on the parameterized model.

Abstract: Methods and corresponding systems that are useful in design and fabrication of kinetic wire mechanisms or characters. The method includes a computational technique for the design of kinetic wire mechanisms tailored for fabrication on consumer-grade hardware such as a desktop CNC bending device. The method takes as input a skeletal animation of the mechanism to be fabricated and estimates, from the skeletal animation, a cable-driven and compliant wire structure, which matches user-selected keyframes. To enable localized deformations, the technique involves shaping the mechanism's body (i.e., the wire) into functional spring-like entities at a set of locations along the length of the mechanism's body.

Abstract: An automated mechanical design analysis system includes a computing platform having a hardware processor and a system memory storing a software code. The hardware processor executes the software code to receive an input model of a mechanical object, identify one or more design parameter(s) of the input model for automated analysis, and perform a parametric mapping of the input model based on the design parameter(s) to produce a parameterized model corresponding to the input model. The hardware processor further executes the software code to embed the parameterized model in a grid to produce model-grid intersections defining multiple subvolumes of the parameterized model, and generate a simulation of the input model based on the model-grid intersections and the subvolumes, where the simulation of the input model provides a differentiable mathematical representation of the input model.

Abstract: A robot control method, and associated robot controllers and robots operating with such methods and controllers, providing computational vibration suppression. Given a desired animation cycle for a robotic system or robot, the control method uses a dynamic simulation of the physical robot, which takes into account the flexible components of the robot, to predict if vibrations will be seen in the physical robot. If vibrations are predicted with the input animation cycle, the control method optimizes the set of motor trajectories to return a set of trajectories that are as close as possible to the artistic or original intent of the provider of the animation cycle, while minimizing unwanted vibration. The new control method or design tool suppresses unwanted vibrations and allows a robot designer to use lighter and/or softer (less stiff) and, therefore, less expensive systems in new robots.

Abstract: A system for providing a user of a virtual reality (VR) system with physical interactions with an object in the real world or in the surrounding physical space while they are concurrently interacting in the virtual world with a corresponding virtual object. The real world object is dynamic with the system including a physical interaction system that includes a robot with a manipulator for moving, positioning, and/or orienting the real world object to move it into contact with the user. For example, the physical object is moved into contact with a tracked body part of the user at a time that is synchronized with a time of an interaction event occurring in the virtual world being created by the VR system. Further, a system is described for providing a dynamic physical interaction to a human participant, e.g., a fast and compelling handover in an augmented reality (AR) system.

Abstract: Methods and corresponding systems that are useful in design and fabrication of kinetic wire mechanisms or characters. The method includes a computational technique for the design of kinetic wire mechanisms tailored for fabrication on consumer-grade hardware such as a desktop CNC bending device. The method takes as input a skeletal animation of the mechanism to be fabricated and estimates, from the skeletal animation, a cable-driven and compliant wire structure, which matches user-selected keyframes. To enable localized deformations, the technique involves shaping the mechanism's body (i.e., the wire) into functional spring-like entities at a set of locations along the length of the mechanism's body.