Abstract: A learning device includes a setting unit configured to set a first value for a parameter of a communication device controlled by a computer using a learned model; a reinforcement learning unit configured to allow a learning model to learn; a model extraction unit configured to extract, as a learned model, the learning model; a model evaluation unit configured to determine whether performance of the learned model has reached first requirement; an updating unit configured to update the first value to a second value when the performance is determined to have reached the first requirement; and a model selection unit. The model evaluation unit determines whether the performance of the learned model updated to the second value satisfies second requirement. When the performance of the learned model updated to the second value is determined to satisfy the second requirement, the model selection unit selects that learned model.

Abstract: A learning system for performing reinforcement learning of a cooperative action by agents includes the agents; and a reward granting unit configured to grant a reward. The reward granting unit performs a step of, in the presence of a target agent to which the reward is to be granted, calculating an evaluation value relating to a cooperative action of other agents as a first evaluation value; a step of, in the absence of the target agent, calculating an evaluation value relating to a cooperative action of the other agents as a second evaluation value; and a step of calculating a difference between the first and second evaluation values as a penalty of the target agent and calculating the reward to be granted to the target agent based on the penalty. The target agent performs learning of the decision-making model based on the reward granted.

Abstract: An unmanned vehicle cooperation system for a person to grasp the operation of an unmanned vehicle and to perform a smooth cooperation action with the unmanned vehicle. The unmanned vehicle cooperation system performs a cooperation action involving human intervention between a plurality of agents including an unmanned vehicle. The unmanned vehicle has a processing unit that performs an action processing of the unmanned vehicle with artificial intelligence using a decision making model established by including a plurality of tasks, the plurality of tasks include a cooperation task for performing the cooperation action, and the processing unit executes a consensus building processing that builds a consensus as to execution of the cooperation task between the agents using a negotiation protocol, and a cooperation action processing that executes the cooperation action of the unmanned vehicle on the basis of the consented cooperation task.

Abstract: A learning device is configured to perform reinforcement learning of a policy of an agent by self-play under a multi-agent environment. The multi-agent environment is an asymmetric environment in which at least one of a type of an action performed by the agent, a type of a state acquired by the agent, and a definition of a reward given to the agent is different, and an evaluation index for learning is given to each of agents. The learning device performs performing learning of a predetermined agent of the agents by using a learning model, acquiring the evaluation index for the learning model of the predetermined agent after learning, comparing the evaluation index for the predetermined agent with the evaluation index for the other agent, and a step of setting the agent having a lower evaluation index as a learning target.

Abstract: A decision-making device (2) comprising: an action selection unit (200) for selecting one of a plurality of actions that can be taken in a first state so that an environment performs the selected action; a state acquisition unit (201) for acquiring a second state indicating the state that follows the execution of the action; a reward acquisition unit (202) for acquiring a reward serving as an indicator for indicating whether or not the second state is desirable; and a storage processing unit (203) whereby experience data in which the first state, the action, the second state, and the reward are associated is stored in a storage unit (21) associated with the action, the storage unit (21) being one of a plurality of storage units.

Abstract: A control device determines an operation scheme of an apparatus on a rule base on the basis of state information indicating the state of the apparatus or the state of an environment in which the apparatus operates, calculates a parameter for determining the content of an operation in the determined operation scheme on the basis of the state information and a learning model constructed on the basis of the operation of the apparatus based on the state information and evaluation of the operation, and commands the apparatus to execute the operation.

Abstract: A movable range of angle of each of the plurality of joints and a safety region defined within the movable range are set. An angle command value is generated to each of the plurality of joints, based on current angle data and a distal end position command value. A fault avoidance control is carried out to make a change rate of the angle command value small, when the angle command value is generated to either of the plurality of joints, and the angle command value of the joint exceeds the safety region.

Abstract: A data transmission apparatus (100) selects transmission object data to a remote control terminal (101) based on acquired point group data (20). At this time, an upper limit of a data quantity of the transmission object data in a predetermined region is determined.

Abstract: An optional position is set as a specification point (10) to a multi-joint manipulator (101). A restrained control command value is calculated to control the multi-joint manipulator in a restrained condition that at least one of degrees of freedom of movement of the multi-joint manipulator is restrained in the specification point, when a control command value to control the multi-joint manipulator is given.

Abstract: A robot (10) has a command predicting section (11) and a command determining section (12). The command predicting section (11) predicts the next second command based on the content of the first command issued from the remote control terminal (101). The command determining section (12) determines whether or not an operation result according to the second command should be transmitted to the remote control terminal (101), based on a comparison result between the second command and a third command issued from the remote control terminal (101) after the first command.

Abstract: Landform data showing the shape of a ground surface (4) is acquired. A specification position is specified on the ground surface. A relative angle (432; 442) between the virtual plane (51) and the ground surface (4) in the at least one inspection point (433; 443) is calculated based on the landform data when the virtual plane (51) set with the reference point (423) and the at least one inspection point (433; 443) is virtually arranged such that the reference point (423) overlaps with the specification position, and such that the virtual plane (51) becomes parallel to the ground surface in the specification position. A landform determination value indicating a flatness of the landform is calculated based on the relative angle (432; 442).

Abstract: An autonomous search system includes an operation terminal 10 and a mobile search device 11. The operation terminal 10 includes an input unit 21 to which a search object and a predetermined action can be separately input, a terminal-side communication unit 24 that transmits, to the mobile search device 11, object data for the input search object and action data for the input predetermined action, and a terminal-side controller 25 connected to the input unit 21 and the terminal-side communication unit 24. The mobile search device 11 autonomously searches for the search object based on the object data transmitted from the operation terminal 10 and carries out the predetermined action relative to the search object based on the action data transmitted from the operation terminal 10.