Abstract: A robotic assistant includes a wheeled base, a storage unit including drawers, a foldable arm connected to a top of the storage unit and including an end of arm tooling (EOAT) connected to a distal end of the foldable arm, an elevation mechanism positioned on the wheeled base and used to move the storage unit up and down, and a control system that receives command instructions. In response to the command instructions, the control system is configured to move the wheeled base, open or close the one or more drawers, actuate movement of the foldable arm and the EOAT to pick up and place external objects from/to a determined location, and control the storage unit to move up/down.

Abstract: A wheeled base includes a housing, two driven wheeled mechanisms positioned on a bottom of the housing and on opposite sides of the housing, at least one passive wheel positioned on the bottom of the housing, actuated feet positioned on the bottom of the housing and configured to move up and down, sensors, and a battery pack arranged within the housing. The two driven wheeled mechanisms each includes a damping mechanism, and each damping mechanism includes at least two dampers configured to absorb impact caused by an upward movement of the housing, and absorb impact caused by a downward movement of the housing.

Abstract: The present disclosure provides a sentence generation method as well as a sentence generation apparatus and a smart device. The method includes: obtaining an input sentence; searching for structurally similar sentence(s) of each input sentence, where the structurally similar sentence(s) are structurally similar to the input sentence; finding semantically similar sentence(s) of the structurally similar sentence(s); parsing the input sentence and the structurally similar sentence(s) to obtain a subject block, a predicate block, and an object block to rewrite the semantically similar sentences to generate a new sentence; filtering the new sentence based on a preset filtering condition; and labeling the filtered new sentence as a semantically similar sentence of the input sentence. In this manner, a plurality of new sentences with different sentence patterns can be generated based on the same input sentence, which improves the controllability in generating the sentences and saves the labor cost therein.

Abstract: The present disclosure provides a robot pose determination method including: collecting laser frames; calculating a current pose of the robot in a map pointed by a first pointer based on the laser frames, and obtaining an amount of the laser frames having been inserted into the map pointed by the first pointer; inserting the laser frames into a map pointed by the first pointer, if less than a first threshold; inserting the laser frames into the map pointed by the first pointer and a map pointed by a second pointer, if greater than or equal to the first threshold and less than a second threshold; and pointing the first pointer to the map pointed by the second pointer, pointing the second pointer to a newly created empty map, and inserting the laser frames into the map pointed by the first pointer, if equal to the second threshold.

Abstract: A linear joint includes a motor assembly includes a rotating shaft for outputting motion; a transmission mechanism including a screw and a nut threadedly connected to the screw, the nut being coaxial with respect to and securely connected to the rotating shaft so as to be rotatable together with the rotating shaft; and a rod connected to a first end of the screw so as to move together with the screw along a lengthwise direction of the screw.

Abstract: Embodiments of the disclosure provide systems and methods for speech detection. The method may include receiving a multichannel audio input that includes a set of audio signals from a set of audio channels in an audio detection array. The method may further include processing the multichannel audio input using a neural network classifier to generate a series of classification results in a series of time windows for the multichannel audio input. The neural network classifier includes a causal temporal convolutional network (TCN) configured to determine a classification result for each time window based on portions of the multichannel audio input n the corresponding time window and one or more time windows before the corresponding time window. The method may additionally include determining whether the multichannel audio input includes one or more speech segments in the series of time windows based on the series of classification results.

Abstract: Embodiments of the disclosure provide methods and systems for implementing a function for an application using a middleware on a computer. An exemplary method may include initializing an application node in the middleware corresponding to the application, the initializing comprising binding the function to the application node and associating the application node with a cryptor, connecting the middleware to a remote device through a transport layer using the cryptor, publishing a message to invoke the function from the application node, through the middleware, to the cryptor, and communicating with the remote device to invoke the function, using the cryptor, based on the message.

Abstract: The present disclosure provides a humanoid robot and its control method and computer readable storage medium. The method includes: obtaining a current torque of a sole of the humanoid robot, an inclination angle of the sole, an inclination angle of a first joint of the humanoid robot, and an inclination angle of a second joint of the humanoid robot; calculating current feedforward angular velocities of motors of the first and second joints through the obtained information; calculating feedback angular velocities of the motors of the first and second joints; and obtaining inclination angles of the joints based on the feedforward angular velocities of the motors and the feedback angular velocities of the motors, and performing, through the motor of the second joint, a deviation control on the joints according to the inclination angles of the joints.

Abstract: The present disclosure provides a method for extracting a face detection image, wherein the method includes: obtaining a plurality of image frames by an image detector, performing a face detection process on each image frame to extract a face area, performing a clarity analysis on the face area of each image frame to obtain a clarity degree of a face, conducting a posture analysis on the face area of each image frame to obtain a face posture angle, generating a comprehensive evaluation index for each image frame in accordance with the clarity degree of the face and the face posture angle of each image frame, and selecting a key frame from the image frames based on the comprehensive evaluation index. Such that the resource occupancy rate during image data processing may be reduced, and the quality of the face detection process may be improved.

Abstract: The present disclosure provides a localization method as well as a robot using the same. The method includes: obtaining laser scan points and particles; mapping each of the laser scan points to a global coordinate system based on each of the particles to obtain global boundary points of each of the particles; finding a matching boundary point in the global boundary points by comparing the global boundary points of the particle with points corresponding to static objects in a known map; calculating a distance between the matching boundary point of the particle and the points corresponding to the static objects, and increasing a weight of the matching boundary point if the distance is less than a preset threshold; calculating a weight of the particle by matching the global boundary points of the particle with the known map; and estimating a localization result.

Abstract: A scene data obtaining method as well as a model training method and a computer readable storage medium using the same are provided. The method includes: building a virtual simulation scene corresponding to an actual scene, where the virtual simulation scene is three-dimensional; determining a view frustum corresponding to preset view angles in the virtual simulation scene; collecting one or more two-dimensional images in the virtual simulation scene and ground truth object data associated with the one or more two-dimensional images using the view frustum corresponding to the preset view angles; and using all the two-dimensional images and the ground truth object data associated with the one or more two-dimensional images as scene data corresponding to the actual scene. In this manner, the data collection does not require manual annotation, and the obtained data can be used for training deep learning-based perceptual models.

Abstract: A robotic leg assembly includes a main body having a first end and an opposite second end, a first servo arranged at the first end of the main body, a second servo connected to the main body and comprising an output shaft, a servo holder arranged at the second end of the main body, a third servo received in the servo holder, and a linkage bar mechanism. Each of a first initial angle of the first rotary member and a second initial angle of the second rotary member is in a predetermined range.

Abstract: A control system for a neck mechanism includes a perception system configured to track movement of an object, and a perception control system that controls a rotary motor to yaw a platform and controls a first linear actuator and a second linear actuator that is in parallel with the first linear actuator to pitch and roll the platform according to a target position of the platform. The perception system tracks movement of the object by estimating its position and pose in 3D space and the platform is moved according to a vision-based position and pose estimation result.

Abstract: A connection assembly for holding a support member in position includes: a base defining a receiving hole to receive an end portion of the support member, and a guiding hole in a lateral surface thereof, which communicates with the receiving hole; a fixing block; a locking member slidably received in the guiding hole and slidable between a first position where the locking member is engaged with the support member so as to hold the support member in position, and a second position where the locking member is disengaged from the support member; and an actuating member connected to the fixing block and the locking member, the actuating member being configured to drive the locking member to move between the first position and the second position.

Abstract: A mobile robot control method includes: acquiring a first image that is captured by a camera on a robot when the robot is in a desired pose; acquiring a second image that is captured by the camera on the robot when the robot is in a current pose; extracting multiple pairs of matching feature points from the first image and the second image, and projecting the extracted feature points onto a virtual unitary sphere to obtain multiple projection feature points, wherein a center of the virtual unitary sphere is coincident with an optical center of coordinates of the camera; acquiring an invariant image feature and a rotation vector feature based on the multiple projection feature points, and controlling the robot to move until the robot is in the desired pose according to the invariant image feature and the rotation vector feature.

Abstract: The present invention discloses a question answering method including obtaining a first question and a first category of the first question, combining the first question with each of preset second questions corresponding to the first category to form question groups, inputting the question groups into a trained deep retrieval matching classification model to obtain a first probability of a first classification label of each of the question groups, inputting the first question into a gradient boosting decision model to obtain a second category of the first question, obtaining a second category of the second questions, adjusting the first probability of the first classification label of each of the question groups, according to the second category of the second questions and the second category of the first question in each of the question groups, and outputting a reply according to adjusted first probabilities for solving a problem of low reply accuracy.

Abstract: The present disclosure provides a robot relocalization method including: obtaining a level feature of an object in a laser map and calculating a first pose list; matching a laser subgraph point cloud collected by the robot with the first pose list to obtain a second pose list, if a distance between the level feature of the object and an initial position of a relocation of the robot is smaller than a threshold; splicing the laser subgraph point cloud into subgraphs, and performing a multi-target template matching to obtain a first matching candidate result; filtering the first matching candidate result based on the second pose list to obtain a second matching candidate result; determining a overlapping area of the second matching candidate result and the subgraph, and matching boundary points in the overlapping area with the laser subgraph point cloud to obtain the result of the relocation of the robot.