Abstract: A vehicle control device (100) includes a recognition unit (132) including an acquisition unit (134) that is configured to acquire information on a traffic management state of signals on a route of a vehicle, the recognition unit being configured to recognize a situation of the surroundings of the vehicle, and a driving control unit (142, 160) that is configured to control acceleration, deceleration, and steering of the vehicle on the basis of the situation of the surroundings of the vehicle recognized by the recognition unit, and in a case that a route along which the vehicle is able to arrive at a destination is able to be set irrespective of which of directions of progression of the vehicle is selected at an intersection at which a signal is provided, the driving control unit is configured to determine a route to the destination on the basis of information acquired by the acquisition unit to cause the vehicle to travel.

Abstract: A vehicle control device includes a recognizer which recognizes objects around a host vehicle, a determiner which determines whether a speed of a preceding vehicle present ahead of the host vehicle in a host lane in which the host vehicle is present among one or more objects recognized by the recognizer is less than a predetermined speed and determines whether predetermined conditions with respect to conditions ahead of the preceding vehicle are satisfied when it is determined that the speed of the preceding vehicle is less than the predetermined speed, and a driving controller which causes the host vehicle to overtake at least the preceding vehicle by controlling the speed and steering of the host vehicle when the determiner determines that the predetermined conditions are satisfied.

Abstract: A vehicle control device (100) includes a recognizer (130) that recognizes a situation occurring near an own-vehicle, and a driving controller (140, 160) that controls acceleration/deceleration and steering of the own-vehicle on the basis of a result of the recognition of the recognizer. When performing travel control according to a behavior of a preceding vehicle that is traveling in front of the own-vehicle such that the own-vehicle follows the preceding vehicle to pass through an intersection, the driving controller does not perform following of the preceding vehicle if a turning start point of the preceding vehicle does not match a turning start point in a route predicted for the own-vehicle to pass through the intersection in a scheduled direction or if a turning angle of the preceding vehicle does not match a turning angle in the predicted route.

Abstract: A vehicle control device (100) of an embodiment includes a moving object recognition unit (131) configured to recognize a moving object that is crossing or is estimated to be about to cross a road on which a vehicle is traveling in a traveling direction of the vehicle, a specifying unit (132) configured to specify an area in which the vehicle is traveling, and avoidance control units (133, 134, 135, 142, 144, and 160) configured to avoid contact with a moving object recognized by the moving object recognition unit by controlling one or both of steering and/or acceleration and deceleration of the vehicle without depending on an operation of a driver of the vehicle, and to change an operation condition of the avoidance control on the basis of an area specified by the specifying unit.

Abstract: A surroundings monitoring device includes a median strip determination unit that determines whether or not there is a median strip in a road around a vehicle, and a progression direction estimation unit that estimates that a plurality of lanes in front of the median strip in a case that viewed from the vehicle among lanes included in a crossroad crossing a road on which the vehicle is traveling are lanes in the same progression direction in a case that the vehicle reaches the crossroad and the median strip determination unit determines that there is the median strip in the crossroad.

Abstract: A vehicle control device includes an other-vehicle recognition unit that recognizes another vehicle stopped near a subject vehicle, a specific-vehicle determination unit that determines whether or not the other vehicle recognized by the other-vehicle recognition unit is a specific vehicle of which a main parking place is the place at which the other vehicle is stopped, and a driving control unit that controls one or both of steering and acceleration/deceleration of the subject vehicle to cause the subject vehicle to travel regardless of an operation of an occupant of the subject vehicle, the driving control unit suppressing, with respect to the other vehicle determined to be the specific vehicle by the specific-vehicle determination unit among the other vehicles recognized by the other-vehicle recognition unit, travel control of the subject vehicle with respect to start of the other vehicle as compared with the other vehicle not determined to be the specific vehicle by the specific-vehicle determination unit.

Abstract: It is possible to perform robot motor learning in a quick and stable manner using a reinforcement learning apparatus including: a first-type environment parameter obtaining unit that obtains a value of one or more first-type environment parameters; a control parameter value calculation unit that calculates a value of one or more control parameters maximizing a reward by using the value of the one or more first-type environment parameters; a control parameter value output unit that outputs the value of the one or more control parameters to the control object; a second-type environment parameter obtaining unit that obtains a value of one or more second-type environment parameters; a virtual external force calculation unit that calculates the virtual external force by using the value of the one or more second-type environment parameters; and a virtual external force output unit that outputs the virtual external force to the control object.

Abstract: A reinforcement learning system (1) of the present invention utilizes a value of a first value gradient function (dV1/dt) in the learning performed by a second learning device (122), namely in evaluating a second reward (r2(t)). The first value gradient function (dV1/dt) is a temporal differential of a first value function (V1) which is defined according to a first reward (r1(t)) obtained from an environment and is served as a learning result given by a first learning device (121). An action policy which should be taken by a robot (R) to execute a task is determined based on the second reward (r2(t)).

Abstract: The present invention provides a motion control system to control a motion of a second motion body, by considering an environment which a human contacts and a motion mode appropriate to the environment, and an environment which a robot actually contacts. The motion mode is learned based on an idea that it is sufficient to learn only a feature part of the motion mode of the human without a necessity to learn the others. Moreover, based on an idea that it is sufficient to reproduce only the feature part of the motion mode of the human without a necessity to reproduce the others, the motion mode of the robot is controlled by using the model obtained from the learning result. Thereby, the motion mode of the robot is controlled by using the motion mode of the human as a prototype without restricting the motion mode thereof more than necessary.

Abstract: It is possible to perform robot motor learning in a quick and stable manner using a reinforcement learning apparatus including: a first-type environment parameter obtaining unit that obtains a value of one or more first-type environment parameters; a control parameter value calculation unit that calculates a value of one or more control parameters maximizing a reward by using the value of the one or more first-type environment parameters; a control parameter value output unit that outputs the value of the one or more control parameters to the control object; a second-type environment parameter obtaining unit that obtains a value of one or more second-type environment parameters; a virtual external force calculation unit that calculates the virtual external force by using the value of the one or more second-type environment parameters; and a virtual external force output unit that outputs the virtual external force to the control object.

Abstract: A reinforcement learning system (1) of the present invention utilizes a value of a first value gradient function (dV1/dt) in the learning performed by a second learning device (122), namely in evaluating a second reward (r2(t)). The first value gradient function (dV1/dt) is a temporal differential of a first value function (V1) which is defined according to a first reward (r1(t)) obtained from an environment and is served as a learning result given by a first learning device (121). An action policy which should be taken by a robot (R) to execute a task is determined based on the second reward (r2(t)).

Abstract: The present invention provides a motion control system control a motion of a second motion body by considering an environment which a human contacts and a motion mode appropriate to the environment, and an environment which a robot actually contacts. The motion mode is learned based on an idea that it is sufficient to learn only a feature part of the motion mode of the human without a necessity to learn the others. Moreover, based on an idea that it is sufficient to reproduce only the feature part of the motion mode of the human without a necessity to reproduce the others, the motion mode of the robot is controlled by using the model obtained from the learning result. Thereby, the motion mode of the robot is controlled by using the motion mode of the human as a prototype without restricting the motion mode thereof more than necessary.

Abstract: It is objective of the invention to provide a highly reliable control apparatus and method which reduces the amount of calculation required for both learning of an input/output relationship and actual control as well as prevent inappropriate outputs from being generated for inputs which have been never learned. According to one aspect of the invention, pairs of an input pattern vector for learning and a target output are distributed to a class based on the target output. Then, a correspondence between each element of the input pattern vector for learning and the target output is learned only in that class, and a distribution function is calculated for distributing a new input pattern vector for learning to that class. After the completion of this learning, the distribution function is used to determine which class a new input pattern vector detected by a sensor belongs to. Finally, an output is calculated according to the learning result of that class.

Abstract: A motion information recognition system takes time-series images of an object in motion and generates image vectors that are analyzed to produce a plurality of eigenvectors that are then stored. Inner product operations are performed between image vectors of a recognized object and a plurality of the stored eigenvectors for obtaining the motion information of the recognized object based on the result of said inner product operations to provide high-speed and easy motion information recognition at reduced computing cost.

Abstract: It is objective of the invention to provide a highly reliable control apparatus and method which reduces the amount of calculation required for both learning of an input/output relationship and actual control as well as prevent inappropriate outputs from being generated for inputs which have been never learned. According to one aspect of the invention, pairs of an input pattern vector for learning and a target output are distributed to a class based on the target output. Then, a correspondence between each element of the input pattern vector for learning and the target output is learned only in that class, and a distribution function is calculated for distributing a new input pattern vector for learning to that class. After the completion of this learning, the distribution function is used to determine which class a new input pattern vector detected by a sensor belongs to. Finally, an output is calculated according to the learning result of that class.

Abstract: The invention provides a high-speed and easy motion information recognition system that is capable of recognizing the motion information directly from the images.