Embodiments are generally directed to generating a training dataset of labelled examples of sensor images and grasp configurations using a set of three-dimensional (3D) models of objects, one or more analytic mechanical representations of either or both of grasp forces and grasp torques, and statistical sampling to model uncertainty in either or both sensing and control. Embodiments can also include using the training dataset to train a function approximator that takes as input a sensor image and returns data that is used to select grasp configurations for a robot grasping or targeting mechanism.

Provided is a grip tool that allows detecting a slip of an object. The grip tool includes a pair of grip portions that sandwiches to grip an object; an opening/closing mechanism that opens and closes the pair of grip portions; and at least one sensor unit disposed in one or both of the pair of grip portions. The sensor unit includes a plurality of force sensors disposed in a predetermined distribution. The plurality of force sensors each have a function of measuring forces in triaxial directions received from the object. The sensor unit can measure a load distribution received from the object. The slip of the object is detectable based on the load distribution.

A robotic system includes a robot having a picking arm to grasp an inventory item and a shuttle. The shuttle includes a platform adapted to receive the inventory item from the picking arm of the robot. The platform is moveable in at least a two-dimensional horizontal plane between a pick-up location located substantially adjacent to the robot and an end location spaced a distance apart from the pick-up location. The system improves efficiency as transportation of the item from the pick-up location to the end location is divided between the robot and the shuttle.

An autonomous system can detect out-of-distribution (OOD) data in robotic grasping systems, based on evaluating image inputs of the robotic grasping systems. Furthermore, the system makes various decisions based on detecting the OOD data, so as to avoid inefficient or hazardous situations or other negative consequences (e.g., damage to products). For example, the system can determine whether a suction-based gripper is optimal for grasping objects in a given scene, based at least in part on determining whether an image defines OOD data.

The invention comprises a system for switching between control points of a robotic system involving an industrial robot including a robot arm with a number of joints and provided with a tool interest point movable in a plurality of degrees of freedom.