109 results about "Robot motion" patented technology

This invention discloses a method for predicting robot motion under time-varying delay conditions. The method determines the number of secondary predictors and their gain parameters based on the maximum time delay, the robot's initial joint position, and the initial joint velocity. The time-varying prediction interval of each secondary predictor is determined by the time-varying delay magnitude. The robot's state from time-delayed motion to the current motion interval is divided into motion sub-states, the same number as the number of secondary predictors. Secondary predictors are constructed, the same number as the number of motion sub-states. The secondary predictors are connected in series, with the time-varying delay robot joint position and velocity signals as input to the first secondary predictor, and the predicted values ​​output by the previous secondary predictor as input to each subsequent secondary predictor. The last secondary predictor outputs the predicted values ​​of the robot's actual joint position and actual joint velocity. This invention offers high prediction accuracy, wide applicability, ease of engineering application and promotion, and contributes to the development of robotics technology.

An apparatus, comprising a memory, configured to store a first data set, comprising kinodynamic data representing a plurality of human-directed movements of a robot; and a second data set, comprising linguistic descriptors of the plurality of human-directed movement of the robot; and a processor, configured to generate a third data set based on the first data set and the second data set, wherein the third data set comprises a plurality of motion primitives of the robot.

This invention relates to the field of single-axis robot technology, and more particularly to an intelligent single-axis robot anomaly monitoring system. The system first acquires mechanism dynamics model data and simultaneously collects real-time monitoring data from the transmission system and the multi-force sensor nodes at the motion execution head. Second, it performs collision identification processing on the data, generating a collision anomaly feature distribution map matching the model's spatial position. Subsequently, combining the map with long-stroke motion constraints, it dynamically and visually reconstructs the dynamics model, outputting an interface with anomaly level markers. Finally, it associates and maps the visualization interface with the real-time data to generate emergency protection and maintenance strategies, triggering a robot protection response through an industrial control port. This invention achieves high-precision positioning and hierarchical visualization of collisions at the motion execution head of a long-stroke single-axis robot, constructs a closed-loop intelligent protection mechanism, and significantly improves equipment safety protection capabilities and operational efficiency.

The invention relates to a foot-type robot cooperative control method based on environment and cluster interaction force, belongs to the technical field of foot-type robot control, and solves one of the problems that an existing foot-type robot cluster is poor in complex terrain adaptability, unstable in dynamic motion control and weak in multi-robot cluster cooperative capability. The method comprises the following steps: obtaining individual state information and neighbor state information of each foot type robot in a cluster at the current moment, and constructing a state space of each foot type robot; based on the state space of each foot type robot, the trained control model of each foot type robot is used to generate each joint torque instruction of each foot type robot at the next moment so as to control each foot type robot to move; wherein the foot type robot control model is a strategy network obtained through reinforcement learning training; during training, a global state space is constructed by aggregating the state spaces of all the foot type robots in the cluster, and the value of action strategies of all the foot type robots in the cluster is evaluated and optimized.

The application discloses a truss environment-oriented two-legged robot crawling landing point sequence planning method and belongs to the technical field of robots; the method first constructs a truss undirected graph topology model based on the spatial connection relationship of truss joints and members, takes nodes to represent the truss joints, takes edges to represent the members and takes the member length as the edge weight; the A star algorithm is adopted on the topology model to search for the node sequence and the member sequence of the robot from the starting member to the target member; subsequently, the axial position of the member is described through a normalized member parameter model, and the non-connected feasible landing area set of the robot foot end is represented in combination with One-Hot coding; on this basis, the least number of steps required for the robot to reach the target position is taken as an optimization target, a mixed integer quadratic constraint programming model is established, the landing point sequence is solved, and the optimal landing point sequence satisfying the truss topology constraint and the robot motion constraint is obtained. The application can realize the landing point sequence planning of a two-legged truss crawling robot on a complex truss structure.

The application discloses a model training method and device and electronic equipment. The method comprises the following steps: acquiring space availability data, embodied visual question answering data and robot demonstration data; constructing a multi-modal training data set based on the space availability data, the embodied visual question answering data and the robot demonstration data; pre-training an initial visual language model using the multi-modal training data set to obtain a target visual language model; adding a motion head to the target visual language model to obtain an initial visual language motion model; screening a robot motion trajectory data set from the robot demonstration data; pre-training the initial visual language motion model using the motion trajectory data set to obtain a target visual language motion model; and the target visual language model in the target visual language motion model is used to obtain a fused semantic representation according to input data, and the motion head in the target visual language motion model is used to output a motion control parameter of a robot according to the fused semantic representation.

This invention discloses a vehicle automation testing method, equipment, and medium based on embodied robots, relating to the field of industrial control technology. The method includes: acquiring a test task description of the target vehicle and performing embodied robot pose calibration, vehicle operable component identification, and association of onboard signal acquisition channels to form a test initialization configuration set; evaluating the differences in control state responses corresponding to the test initialization configuration set under the current vehicle state to generate an adaptive test candidate sequence; and performing convergence control on the execution process of the updated adaptive test candidate sequence based on a confidence convergence determination flag to generate a vehicle automation test report. This invention associates the embodied robot's action execution process with the corresponding onboard signal acquisition data under unified action semantics and time constraints, enabling the vehicle control response to form a quantifiable and continuously updatable state determination basis under different action triggering conditions.

This invention relates to a robot multi-joint motion deviation feedback control system and method based on temporal CNN, belonging to the field of robot motion intelligent control technology. The system includes a multi-source temporal data acquisition module, a temporal data preprocessing module, a temporal CNN spatiotemporal feature extraction module, a multi-dimensional motion deviation calculation module, a hierarchical adaptive feedback control module, a robot motion execution drive module, and a standard motion calibration and storage module. The method includes steps such as standard motion temporal calibration modeling, real-time temporal data acquisition and preprocessing, temporal CNN spatiotemporal feature extraction, multi-dimensional motion deviation calculation, hierarchical feedback command generation, motion closed-loop correction, and cyclic regulation. By employing temporal CNN to extract both spatial and temporal features of the motion, combined with hierarchical feedback and temporal prediction mechanisms, the invention overcomes the shortcomings of traditional control methods, such as feedback lag and inaccurate deviation identification, thereby improving the robot's multi-joint motion control accuracy, response speed, and environmental adaptability.

The application provides a tipping machine connection hooking robot action coordination method and device, relates to the technical field of data processing, and comprises the following steps: firstly, acquiring multi-source real-time data of a tipping machine, a robot end effector and a wrist; based on the data, performing coupling feature learning through a deep network to obtain a high-dimensional hidden space representation of a system coordination state; then, performing explicit decoupling processing on the representation, dynamically isolating interference features generated by tipping machine overturning motion, and extracting core feature representation of hooking action; subsequently, based on the core feature representation, performing energy optimal joint optimization under the premise of satisfying action geometry and dynamics hard constraints to generate a coordinated action sequence; finally, correcting network internal parameters in real time through an online adaptive compensation mechanism to cope with dynamic uncertainties such as cargo sliding, so as to output a final coordinated control instruction. The application can effectively strip strong dynamics interference, and realize high-precision, high-energy-efficiency and robust coordinated operation.

A collaborative robot system. A technique is provided for appropriately controlling the charging period of the collaborative robot. The collaborative robot system includes a robot that works alongside a user in a designated location. The collaborative robot system comprises: a battery that supplies power to enable the robot to move; a display unit that displays the battery's charge level; a storage unit that stores the robot's work content; a calculation unit that calculates the predicted power required for future work content based on the power consumed when performing the work content stored in the storage unit; and a notification unit that notifies the battery of the charging period based on the predicted power.

The application relates to the technical field of robot motion control, and discloses a high-precision control method and system for robot motion and a storage medium. The method comprises the following steps: after receiving an original instruction, performing encoder calibration processing based on the angle position information recorded by an encoder and the target position deviation, generating a basic control signal quantity of the rotation angle of a motor, converting a distance to be walked into a control signal quantity of the rotation angle of left and right wheels through calibrated wheel diameter parameters and wheelbase parameters, and outputting a driving control signal for controlling the rotation of the left and right wheels after superimposing an error compensation control quantity generated through a filtering algorithm. The application improves the motion control precision and self-adaptability of a robot in a dynamic environment.

The present application relates to a kind of fusion ground action force perception's legged robot cluster cooperative control method, belong to robot decision and control technical field, solve the poor adaptability of existing legged robot control method in complex terrain, dynamic motion control is unstable, the problem of weak multi-robot cluster cooperation ability.Method includes: obtaining the state information and environmental perception information of each legged robot and constructing the state space of each legged robot;Based on the state space, using the trained legged robot control model, the motion control instruction of each legged robot in next time is generated to control the motion of each legged robot;Wherein, legged robot control model is the trained actor network;When training the actor network, using critic network, with the global state space constructed by the state space of all legged robots in cluster and the action strategy of all legged robots in cluster as input, the joint value of legged robot action in cluster is evaluated.

This invention provides a motion path planning and obstacle avoidance system for a laser demolition robot, comprising: a calibration and uncertainty module for calculating the extrinsic parameters of the laser head and end effector, the extrinsic parameters of the radar and chassis, and statistical errors; an environment representation module for obtaining an environment representation using multi-source calibration and uncertainty parameters; a risk constraint module for calculating the direct irradiation risk and reflection risk of the main laser beam to obtain risk constraints; a task modeling module for obtaining an optimized cutting task based on the cutting task; a cutting model module for predicting the total cutting depth and constructing a quality predictor; a global coarse planning module for performing graph search to obtain a global candidate access sequence for the chassis; a trajectory optimization module for optimizing the segmented temporal trajectory of the robotic arm to obtain an executable optimized temporal trajectory; and a dynamic obstacle avoidance module for dynamic obstacle avoidance. This invention enables a mobile laser demolition robot to autonomously plan and safely execute high-quality tasks in high-risk and complex environments.

The application discloses a kind of multi-machine cooperation SLAM methods based on active deep reinforcement learning.The method comprises the following steps: running ORB-SLAM2 program to robot, and the initial motion trajectory pose graph of multiple machines is obtained by pose estimation of image acquisition through camera;Based on the obtained robot motion trajectory pose graph, more accurate pose is obtained by using deep reinforcement learning TD3 algorithm training to optimize trajectory;On the basis of reinforcement learning algorithm, active perception strategy is introduced to optimize the pose of multiple machines simultaneously, and according to the real-time SLAM estimation probability value P, the corresponding robot is selected to optimize the pose information by TD3 algorithm;The pose information and actual distance information of each robot are transmitted between robots, and the back-end optimization of SLAM trajectory is carried out using TD3 algorithm, to eliminate the cumulative error effect.The application can effectively eliminate the error accumulation in SLAM system, improve the positioning and mapping accuracy of SLAM, and there is no loop restriction, which increases the robustness of SLAM system.

The application discloses a kind of man-machine cooperation environment under mechanical arm full connecting rod dynamic obstacle avoidance and planning control method, comprising: the joint motion information obtained is input to the motion classification model and is classified, according to classification result, through corresponding joint position prediction network model, human motion prediction is carried out, and three-dimensional joint position of future certain time step is obtained;According to the distance between man-machine sphere and capsule body model, the repulsive field of obstacle is established, and the attractive field is established according to the distance of target point to the end of mechanical arm;Desired angle of each joint is calculated;The desired trajectory of each joint in future time domain is obtained, cost function is constructed, then rolling optimization is carried out, so that the optimal joint angular velocity control sequence is obtained to control robot motion.The application can realize that mechanical arm can effectively complete full connecting rod compliant obstacle avoidance while completing corresponding task, satisfy the real-time and compliance of mechanical arm full connecting rod dynamic obstacle avoidance.

The application provides an entropy-aware robot control method and system based on expert demonstration, and relates to the field of embodied intelligence technology, which comprises obtaining a task instruction input by a user, observation data and ontology perception data of a robot at a current time; cleaning and feature extracting the data to obtain multiple features, and retrieving a reference video in an expert reference video database; encoding and splicing the multiple features to generate a multi-modal feature representation; feature extracting the reference video to obtain reference video features; processing the multi-modal feature representation and the reference video features through a skill generation model to obtain a discrete skill codebook index probability; calculating an information entropy of the probability, and determining a sampling candidate number of a current candidate skill based on the information entropy; generating a skill token sequence based on the discrete skill codebook index probability and the sampling candidate number; and finally decoding the skill token sequence to obtain a robot action sequence. The control method provided by the application can guarantee accurate control of a robot in a complex environment.

The application belongs to the technical field of robot control, and discloses a mobile robot dynamic obstacle avoidance planning system and method based on visual navigation, real-time motion parameters such as obstacle position, speed and acceleration are extracted based on pose semantic perception results, which are input into a social LSTM network together with obstacle semantic types and robot motion parameters, interaction behaviors such as avoidance and following among multiple obstacles are captured through a social pooling layer, the influence of robot motion on obstacle trajectory is quantified through an interaction modeling layer, and a trajectory probability distribution and a prediction confidence of 3-5 time steps are output, while a collision probability, a minimum collision time and a minimum control amount required for collision avoidance are accurately calculated; through a lightweight Transformer double-branch shared network, a pose solution and obstacle semantic segmentation share feature extraction backbones, and robot motion constraints are integrated to realize multi-task parallel output of single-frame data, and perception link delay is shortened.

A robot motion reorientation method, device, equipment and medium are disclosed. The method comprises: determining target rigid body pose information of a reference object in a first calibration posture and target chain link pose information of a robot in a second calibration posture; determining bias calibration information of each target matching item in the target rigid body pose information and the target chain link pose information, respectively; wherein the bias calibration information comprises translation bias calibration information and rotation bias calibration information; determining second motion pose information of the robot according to the first motion pose information of the target object and the bias calibration information of each target matching item; determining target joint configuration information of the robot according to the second motion pose information, so as to control the robot to perform a corresponding action based on the target joint configuration information. The present scheme can improve the calibration efficiency, accuracy and consistency by automatically calibrating the matching item bias of the imitation object rigid body pose and the robot chain link pose in the calibration posture.

The application provides a meat quantitative cutting method and system based on point cloud, and a medium. The method comprises the following steps: obtaining three-dimensional point cloud data of meat; analyzing the point cloud data to determine the main direction of the meat; creating a virtual cutting plane perpendicular to the main direction, and determining a final cutting plane by iteratively adjusting the spatial position of the plane. The iterative adjustment comprises the following steps: calculating the volume of the part to be cut defined by the current virtual cutting plane, comparing the calculated volume with a target volume, and moving the virtual cutting plane according to the comparison result until the absolute value of the difference between the two is less than a preset error threshold. Finally, according to the parameters of the final cutting plane, the robot motion parameters are generated and the robot is controlled to perform cutting. Through the closed-loop feedback mechanism, the application can accurately determine the cutting position that meets the self-defined weight requirement, realizes accurate quantitative cutting, and improves the automation level and raw material utilization rate.

This invention provides a motion adaptive control method and system for a wheel-legged hybrid robot, belonging to the field of robot control technology. The method includes: constructing a multi-mode dynamic model and calibrating sensor parameters; accurately perceiving terrain, obstacles, and ground adhesion characteristics, predicting dynamic obstacle trajectories, and perceiving terrain abrupt changes; presetting multi-parameter fusion switching thresholds for three motion modes: single-wheel, single-leg, and wheel-leg hybrid; formulating dynamic switching decisions; constructing an adaptive control model to perform detailed control of the three motion modes and transition control between modes; and performing multi-source error collaborative compensation for errors in the three motion modes. This invention, through the combination of multi-mode dynamic modeling and cross-modal fusion perception, accurately captures the robot's dynamic characteristics and complex environmental parameters, significantly enhancing the robot's adaptability to diverse terrains and dynamic obstacles such as flat surfaces, slopes, and steps, and enabling rapid identification of terrain abrupt changes and prediction of dynamic obstacle trajectories.

The application discloses a humanoid robot action reorientation control method, which comprises the following steps: obtaining the reorientation motion data sequence of the humanoid robot in high dynamic action according to the target human reference motion data sequence, wherein the reorientation motion data comprises reorientation motion data corresponding to multiple time points respectively; constructing the target body height function corresponding to the humanoid robot; determining the preset high dynamic action trajectory of the humanoid robot in high dynamic action according to the reorientation motion data sequence and the target body height function, and performing filtering processing on the preset high dynamic action trajectory to obtain the target preset high dynamic action trajectory; and controlling the humanoid robot to perform high dynamic action according to the target preset high dynamic action trajectory. The humanoid robot motion data can be mapped to the humanoid robot, the error and local jitter in the process of controlling the humanoid robot to perform high dynamic action can be solved, and effective control of the humanoid robot action reorientation is realized.

This invention discloses a hierarchical reinforcement learning motion planning and control method for legged robots in sparse terrain. It extracts semantically rich terrain features from elevation maps using a multi-scale perceptual encoder. These terrain features, along with the robot's real-time state and target commands, serve as input to a conditional denoising network. This network models landing point planning as a parallel, progressive denoising process, generating a sequence of robot landing points. During this process, an inverse kinematics feasibility verification layer evaluates the kinematic reachability of candidate landing points in real time, thus determining the planned landing point sequence. Based on the planned landing point sequence and the robot's real-time state, robust joint motion commands are output using reinforcement learning to control the legged robot's movements. This invention effectively solves the problems of short-sighted planning, low computational efficiency, and poor robustness in traditional methods in discontinuous, high-elevation, sparse terrain, achieving highly dynamic and stable motion of legged robots in complex environments.

This invention relates to the field of program-controlled robot motion technology, specifically to a motion optimization method, system, and device for amphibious robots. The invention first acquires motion sample information of the amphibious robot in each motion scenario. Then, based on the distribution characteristics and control effect parameters of control parameter combinations at all historical motion moments, it filters out all reference control parameter combinations, thereby determining the distribution reference range of road condition data corresponding to each reference control parameter combination. Finally, based on the distribution reference range of the amphibious robot's current road condition data, it determines the control parameters of the amphibious robot and performs motion control. This invention, based on the effective reference provided by the amphibious robot's historical motion data and control strategies, can adaptively match control strategies based on the current motion situation, avoiding rule base conflicts in different scenarios and improving the motion control effect of the amphibious robot.

This invention relates to a navigation and positioning system and method for a tunnel invert arch inspection robot based on multi-technology fusion, belonging to the fields of tunnel engineering inspection technology and navigation and positioning technology. The system includes a sensor module, a data processing module, a path planning module, a control module, and an execution module. The sensor module collects robot motion parameters and tunnel environment information; the data processing module fuses data from lidar and inertial measurement unit to achieve high-precision positioning; the path planning module generates a full-coverage square wave detection path based on the tunnel structure; the control module uses a model predictive control algorithm combined with real-time positioning information to generate control commands; and the execution module drives the robot to move along the path to complete the invert arch inspection. This invention effectively solves the problems of missing GPS signals and large environmental interference in tunnels, achieving high-precision navigation and positioning with a trajectory tracking error within ±0.06m and a heading angle error within ±1°, thus improving inspection efficiency and comprehensiveness.

The present invention provides a robot motion path generation method and a robot motion path generation apparatus that enable the user to understand the timing in the motion path where the gap between at least one of the robot and the workpiece on the motion path within the derived robot placement area and the placement area is relatively small. [Solution] In this method for generating the operating path of the substrate transport robot 10, the timing 221a in the operating path OP where the size of the gap G between at least one of the robot 10 and the workpiece W on the operating path OP and the placement area 20 is smaller than a predetermined distance is displayed on the display unit 220 in an identifiable manner. This is done using an integrated data ID which is an integrated data ID obtained by combining robot CAD data CD1, placement area CAD data CD2, and workpiece CAD data CD3, used to derive the operating path OP of the substrate transport robot 10 in the placement area 20 where the substrate transport robot 10 and the workpiece W are placed.

The application provides an entropy-aware robot control method and system based on expert demonstration, and relates to the field of embodied intelligence technology, which comprises obtaining a task instruction input by a user, observation data and ontology perception data of a robot at a current time; cleaning and feature extracting the data to obtain multiple features, and retrieving a reference video in an expert reference video database; encoding and splicing the multiple features to generate a multi-modal feature representation; feature extracting the reference video to obtain reference video features; processing the multi-modal feature representation and the reference video features through a skill generation model to obtain a discrete skill codebook index probability; calculating an information entropy of the probability, and determining a sampling candidate number of a current candidate skill based on the information entropy; generating a skill token sequence based on the discrete skill codebook index probability and the sampling candidate number; and finally decoding the skill token sequence to obtain a robot action sequence. The control method provided by the application can guarantee accurate control of a robot in a complex environment.

The invention belongs to the technical field of airplane coating repairing, and particularly relates to an airplane coating precision repairing method based on 3D visual data, which comprises the following steps: shooting a workpiece needing coating repairing to obtain an RGB image, and converting the RGB image into a point cloud picture; after the point cloud picture is processed, a 3D model of a workpiece needing coating repairing is obtained, spraying distance parameters are set, a robot track is generated, path simulation is conducted on the track, if simulation is free of interference, an online laser scanner is arranged to run according to the robot track, secondary detection and positioning are conducted on the surface of the workpiece needing coating repairing, and then the workpiece needing coating repairing is obtained. Acquiring spatial position data of a to-be-restored area; the industrial personal computer generates a robot path middle point according to the spatial position data of the area needing to be sprayed and associates the robot path middle point to an I / O control instruction of the robot; spraying parameters are input into the robot, after a robot motion program is generated, the complete robot track is simulated again, and if simulation is free of interference, the complete robot track is imported into a robot controller to execute spraying.

Production system (1) comprising: a workpiece transfer device (2) that transfers a workpiece (W); a robot (3); and a robot motion device (4) that moves the robot (3), wherein the production system (1) is configured such that, while the workpiece (W) is being transferred by the workpiece transfer device (2), the robot motion device (4) moves the robot (3) and the robot (3) performs an operation while following the workpiece (W), wherein the production system (1) comprises: a determination unit (13) that determines whether a position of the workpiece (W) relative to the robot (3) has exceeded an operating range in which the robot (3) is able to perform work on the workpiece (W) while the robot (3) moves to follow the workpiece (W);wherein the operating area includes a first operating area and a second operating area which is wider than the first operating area and a position compensation unit (14) which, in a case in which the determining unit (13) determines that the position of the workpiece (W) relative to the robot (3) has exceeded the first operating area, performs a first position compensation by controlling or regulating the driving of the workpiece transfer device (2) and / or the driving of the robot motion device (4) such that the position of the workpiece (W) relative to the robot (3) becomes one within the first operating area;and which, in a case where the determining unit (13) determines that the position of the workpiece (W) relative to the robot (3) has exceeded the second operating range, performs a second position compensation with a larger compensation amount than the first position compensation by controlling or regulating the driving of the workpiece transfer device (2) and / or the driving of the robot motion device (4) such that the position of the workpiece (W) relative to the robot (3) becomes one within the first operating range.

The application discloses a kind of flexible production line-oriented multi-robot vision cooperative positioning and grabbing method, it is related to flexible production line robot vision cooperative technical field, comprising: deployment includes industrial camera, 3D laser radar distributed vision node and wrist camera, cover operation area and guarantee field of view overlap and real-time feedback;Multi-source vision data space-time calibration, control synchronization error, unified coordinate and optimize registration accuracy, introduce temperature compensation correction parameter;Cooperative detection tracking workpiece, share results and dynamically adjust tracking mode;Plan global optimal grabbing pose, generate candidate pose and screen;Cooperative control robot motion trajectory, predict avoidance interference;Through vision servo correction pose deviation, trigger grabbing and confirmation, complete transfer.The application improves positioning and grabbing precision and efficiency, enhances multi-robot cooperativity, reduces interference, adapts production line workpiece replacement demand, guarantees long-term operation stability, provides strong technical support for flexible production line.

The application provides a kind of upper limb rehabilitation robot movement intention prediction method and system based on graph-text multimodal fine-grained reasoning, belongs to robot upper limb action recognition technology technical field, by synchronously collecting surface myoelectricity signal of upper limb rehabilitation action and preprocessing, respectively extract one-dimensional time-frequency feature and two-dimensional coding image feature;Construct VGG16-LSTM double branch network training feature extraction model, after extracting fine-grained features, fuse through multi-head cross attention module, then reduce dimension through singular value decomposition, finally input random forest classifier to realize 6 class rehabilitation movement intention prediction, provide accurate movement intention perception ability for upper limb rehabilitation robot.Solve the problem that the accuracy is poor and the robustness is poor when the existing single dimension myoelectricity signal is input into the action prediction model and then the upper limb action intention is predicted.