Methods and apparatus to train interdependent autonomous machines are disclosed. An example method includes performing an action of a first sub-task of a collaborative task with a first collaborative robot in a robotic cell while a second collaborative robot operates in the robotic cell according to a first recorded action of the second collaborative robot, the first recorded action of the second collaborative robot recorded while a second robot controller associated with the second collaborative robot is trained to control the second collaborative robot to perform a second sub-task of the collaborative task, and training a first robot controller associated with the first collaborative robot based at least on a sensing of an interaction of the first collaborative robot with the second collaborative robot while the action of the first sub-task is performed by the first collaborative robot and the second collaborative robot operates according to the first recorded action.

There is provided a recording medium having a program recorded thereon, the program causing a computer to function as: a learning section configured to learn an action model for deciding an action of an action body on a basis of environment information indicating a first environment, and action cost information indicating a cost when the action body takes an action in the first environment; and a decision section configured to decide the action of the action body in the first environment on a basis of the environment information and the action model.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, of training a global policy neural network. One of the methods includes initializing a plurality of instances of the robotic task. For each instance of the robotic task, the method includes generating a trajectory of state-action pairs by selecting actions to be performed by the robotic agent while performing the instance of the robotic task in accordance with current values of the parameters of the global policy neural network, and optimizing a local policy controller that is specific to the instance on the trajectory of state-action pairs for the instance. The method further includes generating training data for the global policy neural network using the local policy controllers, and training the global policy neural network on the training data to adjust the current values of the parameters of the global policy neural network.

A robotic system (new robot) operative for performing at least one task in an environment, the system comprising: learn-from-predecessor functionality governed by a data exchange protocol, which controls short-range wireless knowledge transfer from a short-range wireless transmitter in a predecessor robot system (old robot) to a short-range wireless receiver in said robotic system, said knowledge comprising at least one environment-specific datum previously stored by the predecessor robot.

A robotic system (new robot) operative for performing at least one task in an environment, the system comprising: learn-from-predecessor functionality governed by a data exchange protocol, which controls short-range wireless knowledge transfer from a short-range wireless transmitter in a predecessor robot system (old robot) to a short-range wireless receiver in said robotic system, said knowledge comprising at least one environment-specific datum previously stored by the predecessor robot.

There is provided a recording medium having a program recorded thereon, the program causing a computer to function as: a learning section configured to learn an action model for deciding an action of an action body on a basis of environment information indicating a first environment, and action cost information indicating a cost when the action body takes an action in the first environment; and a decision section configured to decide the action of the action body in the first environment on a basis of the environment information and the action model.

A human collaborative robot system having a function of detecting a force includes a human collaborative robot and a learning unit into which sensing data, internal data, and calculation data are input. The learning unit outputs a first force component applied to the human collaborative robot from outside, a second force component occurring in an operation of the human collaborative robot, and a third force component categorized as noise; and performs learning using supervised data in which inputs and correct labels obtained in advance are collected in pairs, wherein the correct labels of the supervised data are obtained by exerting a force on the human collaborative robot from outside, operating the human collaborative robot over a plurality of paths, and applying noise to the human collaborative robot, and the operation of the human collaborative robot is controlled based on the first force component output from the learning unit.

Methods and apparatus to train interdependent autonomous machines are disclosed. An example method includes performing an action of a first sub-task of a collaborative task with a first collaborative robot in a robotic cell while a second collaborative robot operates in the robotic cell according to a first recorded action of the second collaborative robot, the first recorded action of the second collaborative robot recorded while a second robot controller associated with the second collaborative robot is trained to control the second collaborative robot to perform a second sub-task of the collaborative task, and training a first robot controller associated with the first collaborative robot based at least on a sensing of an interaction of the first collaborative robot with the second collaborative robot while the action of the first sub-task is performed by the first collaborative robot and the second collaborative robot operates according to the first recorded action.

A robotic system (new robot) operative for performing at least one task in an environment, the system comprising: learn-from-predecessor functionality governed by a data exchange protocol, which controls short-range wireless knowledge transfer from a short-range wireless transmitter in a predecessor robot system (old robot) to a short-range wireless receiver in said robotic system, said knowledge comprising at least one environment-specific datum previously stored by the predecessor robot.

A human collaborative robot system having a function of detecting a force includes a human collaborative robot and a learning unit into which sensing data, internal data, and calculation data are input. The learning unit outputs a first force component applied to the human collaborative robot from outside, a second force component occurring in an operation of the human collaborative robot, and a third force component categorized as noise; and performs learning using supervised data in which inputs and correct labels obtained in advance are collected in pairs, wherein the correct labels of the supervised data are obtained by exerting a force on the human collaborative robot from outside, operating the human collaborative robot over a plurality of paths, and applying noise to the human collaborative robot, and the operation of the human collaborative robot is controlled based on the first force component output from the learning unit.