One embodiment can provide a robotic system. The system can include a machine-vision module, a robotic arm comprising an end-effector, a robotic controller configured to control movements of the robotic arm, and an error-compensation module configured to compensate for pose errors of the robotic arm by determining a controller-desired pose corresponding to a camera-instructed pose of the end-effector such that, when the robotic controller controls the movements of the robotic arm based on the controller-desired pose, the end-effector achieves, as observed by the machine-vision module, the camera-instructed pose. The error-compensation module can include a machine learning model configured to output an error matrix that correlates the camera-instructed pose to the controller-desired pose.

A computer-implemented method comprising, receiving data representing a successful trajectory for an insertion task using a robot to insert a connector into a receptacle, performing a parameter optimization process for the robot to perform the insertion task. This parameter optimization includes defining an objective function that measures a similarity of a current trajectory generated with a current set of parameters to the successful trajectory and repeatedly modifying the current set of parameters and evaluating the modified set of parameters according to the objective function until generating a final set of parameters.

A robotic system is disclosed. The system includes a memory that stores a machine learning-based model to provide a scoring function value for a candidate item placement on a pallet on which are plurality of items are to be stacked given a current state value of the pallet and a set of zero or more items placed previously. The system includes one or more processors that use the model to determine a corresponding score for each of a plurality of candidate placements for a next item to be placed and the current state value associated with the current state of the pallet and a set of zero or more items placed previously, select a selected placement based at least in part on the respective scores, control a robotic arm to place the next item according to the selected placement.

A failure prediction device is provided with a machine learning device configured to learn the state of a brake of a motor with respect to data on the brake. The machine learning device observes brake operating state data indicative of an operating state of the brake when the brake is in a normal state, as state variables representative of a current environmental state, and uses the observed state variables to learn a distribution of the state variables with the brake in the normal state.

The present invention provides a method for robot action imitation learning in a three-dimensional space and a system thereof, relates to the technical fields of artificial intelligence and robots. A method based on a series-parallel multi-layer backpropagation (BP) neural network is designed for robot action imitation learning in a three-dimensional space, which applies an imitation learning mechanism to a robot learning system, under the framework of the imitation learning mechanism, to train and learn by transmitting demonstrative information generated from a mechanical arm to the series-parallel multi-layer BP neural network representing a motion strategy. The correspondence between a state characteristic matrix set of the motion and an action characteristic matrix set of the motion is learned, to reproduce the demonstrative action, and generalize the actions and behaviors, so that when facing different tasks, the method does not need to carry out action planning separately, thereby achieving high intelligence.

Systems and methods for detecting an escalator in a surrounding environment by a robotic apparatus are disclosed herein. According to at least one exemplary embodiment, an escalator may be determined based on an escalator detection parameter being met. The escalator detection parameter my further require detection of two side walls separated by a distance equal to a width of an escalator and detection of a depreciation in a floor equal to that observed between a stationary portion and a moving first step of an escalator.

Intelligence guided system and method for fruits and vegetables processing includes a conveyor for carrying produces, various image acquiring and processing hardware and software, water and air jets for cutting and controlling the position and orientation of the produces, and a networking hardware and software, operating in synchronism in an efficient manner to attain speed and accuracy of the produce cutting and high yield and low waste produces processing. The 2nd generation strawberry decalyxing system (AVID2) uniquely utilizes a convolutional neural network (AVIDnet) supporting a discrimination network decision, specifically, on whether a strawberry is to be cut or rejected, and computing a multi-point cutline curvature to be cut along by rapid robotic cutting tool.

A method includes determining a movement of an industrial robot in a manufacturing environment from a first position to a second position. The method also includes displaying an image showing a trajectory of the movement of the robot on a wearable headset. The displaying of the image comprises at least one of: displaying an augmented reality (AR) graphical image or video of the trajectory superimposed on a real-time actual image of the robot, or displaying a virtual reality (VR) graphical image or video showing a graphical representation of the robot together with the trajectory.

A method is for controlling an actuation force exerted by an actuating device having a first working chamber and a second working chamber supplied with pressurized air from a source of pressurized air by a first pressure regulator and a second pressure regulator. The method includes calculating, by an optimization algorithm based on a dynamic model of the actuating device and of the first and second pressure regulators, desired values for control signals for the first and second pressure regulators to generate an actuation force equal to a desired value for the actuation force. An estimated value for the actuation force, estimated values for pressures inside the first and second working chambers and for first derivatives of the pressures, are determined by a state observer based on a measured value for the actuation force and on measured values for the pressures in the first and second working chambers.

A computer-implemented method for performing few-shot imitation is disclosed. The method comprises obtaining at least one set of training data, wherein each set of training data is associated with a task and comprises (i) one of samples of rewards and a reward function, (ii) one of samples of state transitions and a transition distribution, and (iii) a set of first demonstrations, training a policy network embodied in an agent using reinforcement learning by inputting at least one set of first demonstrations of the at least one set of training data into the policy network, and by maximizing a risk measure or an average return over the at least one set of first demonstrations of the at least one set of training data based on respective one or more reward functions or respective samples of rewards, obtaining a set of second demonstrations associated with a new task, and inputting the set of second demonstrations and an observation of a state into the trained policy network for performing the new task.