The present invention relates to systems and methods for accounting for edge cases (i.e. tail data) in automated decision making systems, for example automated robotic picking systems. The systems and methods provide for retraining machine learning (ML) models so that the edge cases can be handled in a manner that requires less (or no) human intervention. The disclosed systems and methods create updated ML models, replacement ML models, and/or supplementary ML models that can provide better performance (e.g. improved automated robotic picking) when edge cases are encountered. Furthermore, the present inventions disclose systems and methods for obtaining training data faster and in a more cost effective manner, which enables the systems and methods disclosed herein to update models at a faster rate, thereby enabling broader, system-wide handling of edge cases in a more effective and efficient manner.

[Solving Means] A production processing apparatus according to the present technology includes a first robot arm and a plurality of first tilt tables. The first robot arm is capable of conveying a work. On each of the plurality of first tilt tables, the work conveyed by the first robot arm can be mounted. The plurality of first tilt tables are tilted a predetermined angle from a horizontal surface at positions on a circumference of a circle with the first robot arm being a center, and the work is subjected to production processing in a state where the work is mounted on one of the plurality of first tilt tables.

A control system includes: a controller configured to operate one or more robots in a real space based on an operation program; and circuitry configured to: operate one or more virtual robots based on the operation program in a virtual space, the one or more virtual robots corresponding to the one or more robots respectively; cause the controller to suspend an operation based on the operation program by the one or more robots; simulate a suspended state of the real space after suspension of the operation by the one or more robots, in the virtual space; and resume at least a part of the operation by the one or more virtual robots based on the operation program, in the virtual space in which the suspended state of the real space has been simulated.

A teaching device includes: a display unit displaying a first icon showing a first attitude of a robot arm, a second icon showing a second attitude of the robot arm, and a first operation unit for performing an operation of designating a third attitude of the robot arm, the first attitude being a state where an angle formed by a first arm and a second arm of the robot arm is a first angle, the second attitude being a state where the angle formed by the first arm and the second arm is a second angle that is different from the first angle, the third attitude being a state where the angle formed by the first arm and the second arm is a third angle equal to or greater than the first angle and equal to or smaller than the second angle; and an operation program generation unit generating the operation program, based on the third attitude designated at the first operation unit.

A robotic system comprises two robots and a control unit. Each robot has a base and an arm extending from the base to an attachment for an instrument. Each arm comprises a plurality of joints whereby the configuration of the arm can be altered. Each robot comprises a driver for each joint configured to drive the joint to move, and position and torque sensors. The control unit controls the drivers in dependence on inputs from the sensors. The control unit: determines the gravitational torques on the joints of the arms of the robots in the arm configurations indicated from the inputs from the position sensors; from the inputs from the torque sensors and the determined gravitational torques, determines residual torques on the joints of the arms of the robots in the indicated arm configurations; calculate a candidate force for each arm which when applied to that arm would cause the determined residual torques; and determines a collision if a candidate force on the arm of the first robot balances an opposing candidate force on the arm of the second robot.

A protocol created in such a format that a series of operations for process targets in the fields of engineering related to living organisms are executable by a robot 10 is acquired (S1). The robot 10 is controlled to implement the operations for the process targets according to the protocol (S2). In order to modify the protocol after the implementation of the operations, modification information on at least one action among basic actions which serve as bases for implementing the operations and is performed on an instrument used by the robot 10 in the operations, and complementary actions which complement the basic actions is acquired (S5). The robot 10 is controlled to produce processed products from the process targets by using the protocol modified based on the modification information (S7).

A protocol created in such a format that a series of operations in cell culture are executable by a robot 10 is acquired (S1). The robot 10 is controlled to implement the operations according to the protocol (S2). In order to modify the protocol after the implementation of the operations, modification information on at least one action among basic actions which serve as bases for implementing the operations and is performed on an instrument used by the robot 10 in the operations, and complementary actions which complement the basic actions is acquired (S5). The robot 10 is controlled to produce cells by using the protocol modified based on the modification information (S7).

A method and a system for manufacturing a mold for creation of complex objects by controlling and moving two end effectors of a robotic system is provided. The two end effectors have a flexible cutting element attached to, and extending between, the two end effectors. The method includes the steps of: defining at least one surface representing the inner surface of the mold; dividing the surface into a number of segments represented by planar curves on the surface; for each planar curve, calculating at least one elastic curve representing the planar curve; for each calculated elastic curve, calculating a set of data corresponding to placement and direction of the two end effectors for configuring the flexible cutting element to a shape corresponding to the calculated elastic curve; and sequentially positioning the end effectors according to each set of data. The flexible cutting element thereby cuts the mold from a block.

The robot includes a first arm having a first hand part provided to a tip end thereof, and at least one joint shaft provided between a pedestal and the first hand part, a first acting part configured to contact a given table-like body on which a plurality of workpieces are able to be placed, while the first acting part is provided to the first hand part, a controller, an imaging unit configured to two-dimensionally image a placement surface of the workpieces in the table-like body in a perpendicular direction to the placement surface, and a recognition part configured to recognize a position of the workpiece by performing a two-dimensional pattern matching based on an image two-dimensionally captured by the imaging unit. The controller vibrates the table-like body by controlling the first arm so that the first acting part acts on the table-like body.

Robotic applications are important in both indoor and outdoor environments. Establishing reliable end-to-end communication among robots in such environments are inevitable. Many real-time challenges in robotic communications are mainly due to the dynamic movement of robots, battery constraints, absence of Global Position System (GPS), etc. Systems and methods of the present disclosure provide assisted link prediction (ALP) protocol for communication between robots that resolves real-time challenges link ambiguity, prediction accuracy, improving Packet Reception Ratio (PRR) and reducing energy consumption in-terms of lesser retransmissions by computing link matrix between robots and determining status of a Collaborative Robotic based Link Prediction (CRLP) link prediction based on a comparison of link matrix value with a predefined covariance link matrix threshold. Based on determined status, robots either transmit or receive packet, and the predefined covariance link matrix threshold is dynamically updated. If the link to be predicted is unavailable, the system resolves ambiguity thereby enabling communication between robots.