An adaptive robot grasp planning technique for bin picking. Workpieces in a bin having random positions and poses are to be grasped by a robot and placed in a goal position and pose. The workpiece shape is analyzed to identify a plurality of robust grasp options, each grasp option having a position and orientation. The workpiece shape is also analyzed to determine a plurality of stable intermediate poses. Each individual workpiece in the bin is evaluated to identity a set of feasible grasps, and the workpiece is moved to the goal pose if such direct movement is possible. If direct movement is not possible, a search problem is formulated, where each stable intermediate pose is a node. The search problem is solved by evaluating the feasibility and optimality of each link between nodes. Feasibility of each link is evaluated in terms of collision avoidance constraints and robot joint motion constraints.

A method of setting a target force upper limit for a robot gripping an object with a gripping unit and operating by force control to bring an acting force close to a target force, includes gripping the object with the gripping unit, performing a pressing operation to press the object gripped by the gripping unit against a contact surface by the force control, performing a pressing force acquisition operation to acquire the force acting on the gripping unit during the pressing operation as a pressing force, repeating a setting change operation to increase the target force, the pressing operation, and the pressing force acquisition operation until a state in which the pressing force is not equal to or larger than the target force appears, and setting a target force upper limit based on the pressing force acquired in the pressing force acquisition operation at a time when the state appears.

A grasping device includes: a grasping part module including a first surface and a second surface and configured to grasp an object between the first surface and the second surface; an arm part configured to change a position of the grasping part module; an imaging unit provided at a position that moves together with the grasping part module and configured to capture an image of at least a part of the object; and a control unit configured to control, based on specified amount information indicating a contact state in a case where a specified amount of the object and the first surface are in contact with each other, and information indicating a contact state captured by the imaging unit, at least one of the grasping part module and the arm part such that an amount of the object that is grasped approaches the specified amount.

Deep machine learning methods and apparatus, some of which are related to determining a grasp outcome prediction for a candidate grasp pose of an end effector of a robot. Some implementations are directed to training and utilization of both a geometry network and a grasp outcome prediction network. The trained geometry network can be utilized to generate, based on two-dimensional or two-and-a-half-dimensional image(s), geometry output(s) that are: geometry-aware, and that represent (e.g., high-dimensionally) three-dimensional features captured by the image(s). In some implementations, the geometry output(s) include at least an encoding that is generated based on a trained encoding neural network trained to generate encodings that represent three-dimensional features (e.g., shape). The trained grasp outcome prediction network can be utilized to generate, based on applying the geometry output(s) and additional data as input(s) to the network, a grasp outcome prediction for a candidate grasp pose.

A robot hand includes at least one finger. The finger includes a first member configured to come into contact with an object to apply a gripping force to the object, and a second member movable with respect to the first member to come into contact with the object.

A work transfer system includes a robot having a hand which holds a workpiece and a sensor which can detect external force acting on the hand, a balancer connected to the hand and can generate lifting force for lifting the hand in a vertically upward direction, a shape measuring device which conducts measuring of a shape of the workpiece, and a controller controlling the robot and the balancer based on the shape of the workpiece, and the controller adjusts a holding position of the workpiece by the hand based on the shape, and controls the lifting force so that an absolute value of the external force in the vertical direction detected by the sensor becomes equal to or smaller than a predetermined first threshold when the workpiece is held at the adjusted holding position and lifted.

A picking system is provided, which is capable of picking up an object even when the object is not registered in advance. The picking system includes: a picking device holding the object; an RGB-D camera acquiring three-dimensional point cloud data of the object to be picked up by the picking device; and a control device controlling the picking device based on a detection result by the RGB-D camera. The control device generates a geometric model of the object by combining simple geometric primitives while referring to the three-dimensional point cloud data, and calculates a holding position of the object for the picking device based on the geometric model.

Deep machine learning methods and apparatus related to manipulation of an object by an end effector of a robot. Some implementations relate to training a deep neural network to predict a measure that candidate motion data for an end effector of a robot will result in a successful grasp of one or more objects by the end effector. Some implementations are directed to utilization of the trained deep neural network to servo a grasping end effector of a robot to achieve a successful grasp of an object by the grasping end effector. For example, the trained deep neural network may be utilized in the iterative updating of motion control commands for one or more actuators of a robot that control the pose of a grasping end effector of the robot, and to determine when to generate grasping control commands to effectuate an attempted grasp by the grasping end effector.

A gripper, an apparatus, and a method for assembling kits of sanitary products in an automated fashion are disclosed. The sanitary products are pre-loaded into containers including a plurality of independent housings for the sanitary products, each of the housings being configured to allow loading of a sanitary product therein and withdrawal of the sanitary product therefrom independently of the other housings, and the kit is assembled in batches on the gripper, and released by the latter into a package.

Methods of performing a plurality of operations within a region of a part utilizing an end effector of a robot and robots that perform the methods are disclosed herein. The methods include collecting a spatial representation of the part and aligning a predetermined raster scan pattern for movement of the end effector relative to the part with the spatial representation of the part. The methods also include defining a plurality of normality vectors for the part at a plurality of predetermined operation locations for operation of the end effector. The methods further include moving the end effector relative to the part and along the predetermined raster scan pattern. The methods also include orienting the end effector such that an operation device of the end effector faces toward each operation location along a corresponding normality vector and executing a corresponding operation of the plurality of operations with the operation device.