Honda Motor Co., Ltd.

Katsu Yamane

Ryusuke Ishizaki

Mirza Saquib Sarwar

Shih-Yun Lo

Allison Mariko Okamura

Soshi Iba

Kenichiro Sugiyama

Rana Soltani Zarrin

Xingwei Wu

Peter Trautman

Joseph Campbell

Chihiro Kurosu

Cara M Nunez

Brian H Do

Akshit Kaplish

Chao Cao

Yalun Wen

A controller for an inverse optimal control approach robot may control movement of a robot body toward a human target along a trajectory according to a cost function. The cost function may include many terms. A first term may be associated with a duration of the trajectory for the robot. A second term may be associated with a social force and a final distance between the robot and the human target. A third term may be associated with a lateral acceleration for the robot. A fourth term may be associated with an angular acceleration for the robot. A fifth term may be associated with a longitudinal acceleration for the robot. A sixth term may be associated with a reduction of centrifugal force applied to the robot body.

Inverse optimal control for human approach

A robot for object manipulation may include sensors, a robot appendage, actuators configured to drive joints of the robot appendage, a planner, and a controller. Object path planning may include determining poses. Object trajectory optimization may include assigning a set of timestamps to the poses, optimizing a cost function based on an inverse kinematic (IK) error, a difference between an estimated required wrench and an actual wrench, and a grasp efficiency, and generating a reference object trajectory based on the optimized cost function. Grasp sequence planning may be model-based or deep reinforcement learning (DRL) policy based. The controller may implement the reference object trajectory and the grasp sequence via the robot appendage and actuators.

Object manipulation

A system and method for social-physical haptic interaction that include receiving sensor data from a plurality of sensors that are disposed within a haptic vest. The system and method also include training a neural network with at least one haptic interaction profile based on the sensor data. The system and method additionally include analyzing at least one haptic interaction profile during execution of at least one haptic application. The system and method further include electronically controlling the haptic vest to provide haptic feedback during the execution of at least one haptic application that is based on at least one haptic interaction profile.

System and method for social-physical haptic interaction

A system and method for fabricating soft sensors that conform to arbitrary smooth geometries that include fabricating a top stretchable layer that includes a set of electrodes of soft sensors that are made of an elastic material. The system and method also include fabricating a bottom flexible layer that is composed of a thin sheet of suitable metal that is patterned using photolithography.

System and method for fabricating soft sensors that conform to arbitrary smooth geometries

A system and method for implementing pedestrian avoidance strategies for a mobile robot that include receiving position data of a pedestrian and the mobile robot from systems of the mobile robot and estimating positions of the pedestrian and the mobile robot based on the position data. The system and method also include determining an expected intersection point of paths of the pedestrian and the mobile robot and an estimated time for the pedestrian to reach and cross the expected intersection point of the paths. The system and method further include implementing a pedestrian avoidance strategy based on the positions of the pedestrian and the mobile robot and the expected point in time when the pedestrian will reach and cross the expected intersection point of the paths.

System and method for implementing pedestrian avoidance strategies for a mobile robot

A system and method for fabricating soft sensors that conform to arbitrary smooth geometries that include fabricating a top stretchable layer that includes a set of electrodes of soft sensors that are made of an elastic material. The system and method also include fabricating a bottom flexible layer that is composed of a thin sheet of suitable metal that is patterned using photolithography. The system and method further include bonding the top stretchable layer to the bottom flexible layer to form a sensor substrate.

An electronic apparatus for a database construction and control of a robotic manipulator is provided. The electronic apparatus stores information associated with a task of a robotic manipulator. The electronic apparatus further receives a first plurality of signals from a first plurality of sensors associated with a wearable device. The electronic apparatus further applies a predefined model on a first set of signals of the first plurality of signals. The electronic apparatus further determines arrow direction information based on the application of the predefined model on the first set of signals. The electronic apparatus further aggregates the determined arrow direction information with information about the first set of signals to generate output information. The electronic apparatus further stores the generated output information for each of a first plurality of poses performed for the task using the wearable device.

Database construction for control of robotic manipulator

A proximity and three-axis force sensor based sensor may include a first taxel including a first electrode formed within a top layer configured in a serpentine pattern, a second electrode formed within a bottom layer, and a dielectric layer positioned between the top layer and the bottom layer and a second taxel including a first electrode formed within the top layer and having a first surface area, a second electrode formed within the bottom layer and having a second surface area, and a ground electrode formed within the top layer above the first electrode of the second taxel having a surface area greater than the first surface area of the first electrode of the second taxel. The second surface area may be different than the first surface area. A first edge of the first electrode may be vertically aligned with a first edge of the second electrode.

Proximity and three-axis force sensor

A robot for physical human-robot interaction may include a number of sensors, a processor, a controller, an actuator, and a joint. The sensors may receive a corresponding number of sensor measurements. The processor may reduce a dimensionality of the number of sensor measurements based on temporal sparsity associated with the number of sensors and spatial sparsity associated with the number of sensors and generate an updated sensor measurement dataset. The processor may receive an action associated with a human involved in pHRI with the robot. The processor may generate a response for the robot based on the updated sensor measurement dataset and the action. The controller may implement the response via an actuator within a joint of the robot.

Physical human-robot interaction (pHRI)

Controlling a robot may be performed by calculating a desired velocity based on an operator contact force and a robot contact force, calculating a transformation matrix based on an operator pose, a robot pose, a robot trajectory, the operator contact force, and the robot contact force, calculating a least square solution to J{circumflex over (v)}based on the desired velocity, calculating a scaling factor based on a current joint position of a joint of the robot, calculating a first trajectory based on the least square solution, the scaling factor, and the transformation matrix, calculating the robot trajectory, calculating a joint command based on the robot trajectory and the robot pose, and implementing the joint command.

Motion retargeting control for human-robot interaction

According to one aspect, a control system for providing stable balance control may include an Hcontroller, a state-feedback circuit, a first feedback loop, and a second feedback loop. The control system may be implemented in a robot as a controller for the robot. The Hcontroller may receive an input signal and generate a control effort signal. The state-feedback circuit may receive the control effort signal as an input and generate an output signal. The feedback loop may include the Hcontroller and the state-feedback circuit and may transfer the output signal of the state-feedback circuit back to the input of the Hcontroller and input a tracking error input signal to the Hcontroller. The tracking error input signal may be the difference between the output signal of the state-feedback circuit and the input signal.

Stable balance controller

A control system and method for generalized admittance control (AC) is provided. The control system includes circuitry communicatively coupled to a robotic manipulator. The circuitry receives, from a sensor system of the robotic manipulator, contact force measurements associated with a physical interaction of the robotic manipulator with an object. The circuitry determines a surface portion of the robotic manipulator on which the physical interaction with the object occurs and samples a set of interaction points from the surface portion. The circuitry computes a generalized velocity for AC of the robotic manipulator by minimizing a cost function. The cost function includes a relationship between the contact force measurements and an approximation term which models application of an actual contact force of the physical interaction on the sampled set of interaction points. The circuitry generates a set of joint control instructions based on generalized velocity to control motion of the robotic manipulator.

Generalized admittance control for robotic manipulators

Embodiments, systems, and methods for navigational planning of a mobile programmable agent are provided. In some embodiments, the navigational planning may include identifying a plurality of dynamic objects in a physical environment having an origin and a destination. The physical environment is divided into a plurality of plane figures. The location of a centroid of each plane figure can then be calculated. A network of segments is formed from the origin to the destination intersecting the centroids. At least one channel is determined from the origin to the destination using a set of segments. A set of gates is identified along the at least one channel. The state of the gates is selectively determined based on movement of the dynamic objects. A pathway can then be identified within the channel for the mobile programmable agent to traverse from the origin to the destination based on the state of the gates.

Growing community of inventors

Mountain View, CA, United States of America

Katsu Yamane