302 results about "Autonomous robot" patented technology

A method for operating a robot, including: capturing images of a workspace; capturing data indicative of movement of the robot; capturing LIDAR data as the robot moves within the workspace; generating a map of the workspace based on the LIDAR data; actuating the robot to drive; discriminating between an object on a floor surface along a path of the robot and the floor surface based on the captured images; actuating the robot to drive until determining all areas of the workspace are discovered and included in the map; and executing a cleaning function.

Some aspects include a method for operating an autonomous robot, including: capturing, with a first sensor disposed on the robot, data of an environment of the robot; generating, with the processor, a map of the environment based on at least the data of the environment; localizing, with the processor, the robot within the environment; capturing, with a second sensor disposed on the robot, data of a floor surface; determining, with the processor, a floor type of areas of the environment based on the data of the floor surface; and determining, with the processor, settings of the robot based on at least the floor type of the floor surface, wherein the settings comprise at least an elevation of each of at least one component of the robot from the floor surface.

The invention discloses an ocean induced polarization and natural potential combined acquisition device and a data processing method, and solves the technical problem that the sulfide ore form cannot be accurately represented in the prior art. The system comprises a first underwater autonomous robot and a second underwater autonomous robot, the first underwater autonomous robot is provided with an induced polarization emission device, and the second underwater autonomous robot is provided with a natural potential acquisition device; the induced polarization emission device comprises an induced polarization control bin, a first watertight cable, an emission end positive pole and an emission end negative pole, and the natural potential acquisition equipment comprises a natural potential acquisition bin, a second watertight cable and a plurality of pairs of Ag / AgCl non-polarized electrodes. According to the method, the distribution characteristics of submarine sulfides can be more accurately explained, the multiplicity of solutions is reduced, and the exploration accuracy and reliability are improved.

The invention relates to the technical field of tunnel defect monitoring, in particular to a tunnel monitoring system based on an autonomous robot, which realizes autonomous inspection and multi-modal data acquisition in a tunnel through a multi-module collaborative autonomous robot system and an integrated image acquisition and ground penetrating radar device, and improves the defect detection efficiency and continuity. Compared with a traditional mode depending on image recognition, a radar scanning image and a representation image are fused, feature extraction and defect reasoning are carried out through a deep learning model, the type and the extension area of the structural hidden defect are recognized, and the recognition capacity of deep structure abnormity is enhanced. Meanwhile, in combination with defect types and extension prediction, the robot is guided to execute supplementary scanning, and data integrity is ensured. Furthermore, cause classification analysis is carried out through a plurality of complete defect detection data, intelligent closed-loop monitoring from defect identification to cause diagnosis is realized, the monitoring cost is effectively reduced, and the accuracy and automation level of cause judgment of tunnel structure diseases are improved.

Some aspects include a method for operating an autonomous robot, including: capturing, with a first sensor disposed on the robot, data of an environment of the robot; generating, with the processor, a map of the environment based on at least the data of the environment; localizing, with the processor, the robot within the environment; capturing, with a second sensor disposed on the robot, data of a floor surface; determining, with the processor, a floor type of areas of the environment based on the data of the floor surface; and determining, with the processor, settings of the robot based on at least the floor type of the floor surface, wherein the settings comprise at least an elevation of each of at least one component of the robot from the floor surface.

The invention relates to a multi-task operation planning method and system for an autonomous robot driven by a body cognition large model. The method comprises the steps that S1, coding is carried out based on an RGB image and a depth image which are obtained in real time, and body visual representation is obtained; s2, acquiring a natural language instruction, performing cross-modal fusion on the natural language instruction and the body visual representation to obtain a fusion feature, and performing multi-task decomposition on the basis of the fusion feature; s3, on the basis of a multi-task decomposition scheme, a diffusion strategy is utilized to generate a continuous action track of a robot end effector; and S4, acquiring a second RGB image and a second depth map after the robot executes according to the continuous action track, and taking the second RGB image and the second depth map as closed-loop feedback signals. The system is used for realizing the method. Compared with the prior art, the multi-task autonomous planning and accurate execution capability of the robot in a dynamic complex environment is remarkably improved by deeply fusing vision and language modalities based on the biont cognition large model and predicting an accurate action trajectory in combination with the diffusion strategy action decision module.

A method for greatly reducing data storage requirements of autonomous robots capable of product inventory is described. Instead using high resolution images for label identification and position tracking, a lower resolution map can be searched. Human or machines can identify position of labels, and via a reverse mapping, the corresponding position of a label on the high resolution image can be identified. Except for those portions of the high resolution image showing a label, most of the high resolution image can be discarded. Further processing on the limited image subset can be used to read the labels.

A system for generating improved notifications relating to events is disclosed. The system includes a sensing network including an event sensor configured to sense events. The system further includes a sensed-event processing system configured to append meta data to information associated with a sensed event. The system further includes a human-machine interface and at least one communication unit configured to communicate amongst the human-machine interface, the sensing network, and the sensed event processing system, the information communicated from the sensing device via the at least one communication unit. The system further includes a notification generation system configured to generate a notification to a user of the human-machine interface, the notification content including at least one user-selectable action option. The communication unit configured to communicate the notification via the human-machine interface and select the notification content, the notification type, or the means of notification based on the meta data.

The invention relates to a super-extensible underwater robot cluster system based on natural spinning and implicit cooperation and a coordination method of the super-extensible underwater robot cluster system, and aims to solve the challenges of a traditional underwater robot cluster system in the aspects of communication, expansibility and adaptability through a simple design and an innovative cooperation mechanism. The system comprises a plurality of autonomous robot monomers, cooperates with adjacent robots through a visual guidance form information interaction mechanism, and transmits state information by using an optical beacon system to realize implicit cooperation and self-organization. According to the method, a coordination mechanism based on consensus initiative is adopted, the robot adjusts behaviors only through local sensing and does not need to depend on global positioning or a complex communication network, the system has high self-healing capacity, and if an individual loses efficacy, other robots can automatically recover cluster stability and task execution. The underwater robot cluster system is low in cost and high in efficiency, and a brand new solution is provided for practical application of large-scale autonomous underwater robot clusters.

The invention discloses a robot automatic obstacle avoidance target recognition system and method, and relates to the technical field of autonomous navigation, and the method comprises the steps: obtaining and fusing an obstacle boundary map, a passable region mask and a dynamic target hot area map, and outputting an environment feature map; inputting the environment feature map into a pre-trained deep reinforcement learning model, and generating an obstacle avoidance instruction set according to the real-time state of the robot and a global path planning result; and acquiring actual path data of the robot, comparing the actual path data with an expected path point by point, and if the continuous deviation feature exceeds a preset continuous deviation feature threshold, inputting the environment feature map, the obstacle avoidance instruction set and the continuous deviation feature into the deep reinforcement learning model for updating. According to the invention, the obstacle avoidance capability of the autonomous robot in a dynamic complex environment is improved through multi-modal sensing fusion and a closed-loop dynamic optimization mechanism.

The invention relates to a visual inertial fusion positioning method and system based on feature optimization, and belongs to the technical field of visual inertial navigation of autonomous robots. According to the method and the system, on the basis of a Geometric Feature Track Selection Visual Native Navigation System (GFT-VINS), through a feature trajectory selection strategy based on normal epipolar geometry, high-quality feature trajectories are screened out, low-quality trajectory interference is avoided, the positioning precision is improved, error accumulation caused by environmental changes and feature anomalies is reduced, and the system stability is enhanced. Meanwhile, in combination with a multi-state Constraint Kalman filter (MSCKF) framework, the efficient calculation capability is kept, the requirements of the autonomous robot for high precision, high stability and efficient calculation in a complex environment are met, and wide application of the autonomous robot is promoted.

Systems and methods described herein relate to generating perception data. In one embodiment, a system extracts, in an offline processing environment, first features from a time sequence of perceptual sensor data to generate a first set of bird's-eye-view (BEV) feature images. The system also extracts second features from the first set of BEV feature images using a BEV feature extractor that performs bidirectional feature-level temporal aggregation to generate a second set of BEV feature images. The system also consumes the second set of BEV feature images using one or more neural-network heads to perform one of: (1) generating automatically labeled perception data to train one or more of an online perception model, an online prediction model, and an online planning model used to control an autonomous robot and (2) validating the performance of an online autonomous stack used to control an autonomous robot.

A robot according to the present invention is mounted on a vehicle which is capable of autonomous travelling, said robot comprising: an arm, orientation and the position of the tip end of which can be discretionarily changed; a vision sensor which is attached to the tip end of the arm; a detection unit which detects the amount of deviation of the vehicle from a specified path on the basis of a marker included in an image that has been captured by the vision sensor; and a travel instruction unit which, on the basis of the amount of deviation, instructs the vehicle to correct the travelling position or the travelling orientation of the vehicle.

The invention provides a robot operation method and system based on a visual language model and reinforcement learning, and relates to the technical field of robot operation, a mechanical arm of a robot operates a target object according to task description, and the robot operation method comprises the steps that to-be-executed task description, a real-time task observation image and a real-time environment observation image are acquired; based on the task description and the environment observation image, decomposing the task description into executable subtask sequences represented by semantic key points by utilizing a visual language model; performing space mapping on the executable subtask sequence; and generating a reference trajectory by taking the mapped coordinates and directions as targets, and finely adjusting the reference trajectory by using the task observation image to obtain a final task operation trajectory. According to the method, robustness is shown in a long sequence task, diversified objects can be flexibly operated, the potential of combination of multi-modal reasoning and self-adaptive control learning is fully excavated, and a new path is developed for constructing an autonomous robot system with a higher capability level.

Systems and methods for estimating performance of an autonomous robot system that is configured to fulfil multiple line orders is described herein. The autonomous robot system comprises a plurality of autonomous guided vehicles configured to transport one or more cases within an environment so as to fulfil the multiple line orders. In some variations, a method includes obtaining first input data, generating a simulation model based at least in part on the first input data, determining a travel duration for each of the plurality of autonomous guided vehicles based on an execution of the simulation model, generating an analytical model based at least in part on the travel duration, and estimating the performance of the autonomous robot system based on an execution of the analytical model. The first input data includes data associated with operation of each of the plurality of autonomous guided vehicles and data representing a layout of the environment.

The invention discloses a low-carbon post-disaster rescue system based on quantum computing and an autonomous robot technology, and belongs to the technical field of post-disaster rescue. The system comprises a post-disaster information acquisition module, a data acquisition and analysis module, a quantum computing decision optimization and resource scheduling module, an autonomous robot task scheduling and execution module and a low-carbon technology model module. Performing ultra-high-speed processing and multi-target optimization decision on post-disaster environment data by adopting a mixed quantum annealing algorithm, and generating a resource scheduling and path planning scheme; the autonomous robot executes post-disaster building evaluation, material delivery and personnel search and rescue tasks based on an AI algorithm, so that the manual risk is reduced; and by combining low-energy-consumption hardware, renewable energy power supply and real-time carbon emission monitoring technologies, the carbon emission in the rescue process is reduced by more than 30%. The intelligent low-carbon rescue system solves the problems of low rescue efficiency, poor safety and high carbon emission in the prior art, is suitable for extreme post-disaster scenes with communication interruption, and provides an innovative solution for intelligent low-carbon rescue.

Thus, there is disclosed an autonomous robotic construction system comprising at least one frame comprising at least one substantially horizontal track and at least one set of substantially vertical track; at least one nozzle; at least one movable device coupled to the frame and configured to move said nozzle around said frame; at least one pump configured to pump building materials through said at least one nozzle; and at least one computer system comprising at least one microprocessor configured to control said at least one nozzle, said at least one movable device and said at least one pump to control deposition of any building materials on a site.

The invention relates to the field of robot technology and artificial intelligence, and discloses a robot fine operation method and system based on a visual language model and semantic key point characterization, a closed-loop multi-body Agent system is constructed, GPT-4o and Ground DINO are combined to accurately understand and disambiguate a natural language instruction of a user, an innovative bridging layer is introduced, and the robot fine operation method and system based on visual language model and semantic key point characterization are obtained. Abstract decision information is converted into bottom execution parameters represented by three-dimensional semantic key points, the problem of decision transfer distortion is solved, meanwhile, a KMP algorithm is improved, and the stiffness and contact force of trajectory tracking are dynamically optimized by introducing a learning-based time-varying hybrid impedance / force control strategy, so that the accuracy of trajectory tracking is improved. According to the method, the generalization success rate of fine operation tasks of the robot and the stability of physical interaction are remarkably improved, seamless connection from high-level semantic understanding to low-level action generation is achieved, and an efficient and reliable technical scheme is provided for autonomous operation of the robot in a complex scene.

An autonomous robot drive assembly includes a force sensing assembly. The force sensing assembly is a force sensing handlebar that is mounted in a specific orientation to allow for a user to manipulate the robot. The handlebar is configured to allow a user to manipulate the handlebar by providing force to the handlebar to move the handlebar from a neutral position. The manipulation of the handlebar causes instructions to be determined for operation of the robot. Based on the manipulation of the handlebar, a drive assembly of the robot moves the robot in accordance with the instructions.

The invention discloses a field live-line safety protection method and system for an autonomous robot, and belongs to the technical field of intelligent live-line detection. The method comprises the following steps: based on a preset adaptive access technology, in a simulation environment, training a virtual simulation model by using a constructed deep reinforcement learning algorithm to complete a specified action so as to train an action algorithm of the virtual simulation model, and transplanting the trained action algorithm to an autonomous robot; based on a preset comprehensive evaluation model, identifying the monitoring data, determining behaviors of the autonomous robot, and generating a sensing result; and generating an electronic map of risk score quantitative distribution based on a perception result, collecting fine-grained action information of the autonomous robot based on the electronic map, and performing active behavior risk early warning based on the fine-grained action information. According to the invention, the performance detection work of the combined transformer can be carried out under the electrified condition, and the power failure loss is reduced.

Systems and methods for performing a task in an operation environment of an aerial device with an autonomous or semi-autonomous robot are described. In some embodiments, a robot is disposed at an end of a boom of an aerial device. The robot may comprise cameras, actuators, sensors, processors, and manipulators that work together to perform tasks fully autonomously or semi-autonomously. Furthermore, the robot may comprise tools for performing the tasks and computer-executable instructions for performing the tasks may be based on the various sensory inputs, the tools, and the tasks to be performed.

Systems and methods described herein relate to detecting and tracking objects. In one embodiment, a system extracts first features from time-sequential perceptual sensor data to generate a first set of bird's-eye-view (BEV) feature images. The system also extracts second features from the first set of BEV feature images using a three-dimensional (3D) detection backbone to generate a second set of BEV feature images. The system also consumes the second set of BEV feature images using a neural-network 3D detection head that is trained with a similarity objective to support an object tracker for use in one of (1) controlling an autonomous robot and (2) generating automatically labeled perception data to train one or more of an online perception model, an online prediction model, and an online planning model used to control an autonomous robot.

In an embodiment of the invention, a sensor suite for an autonomous robotic kitchen system includes a thermometer, a sensor suite housing, a sensing chamber within the sensor suite housing having an intake port and an output port, at least one fan configured to control flow of gas through the sensing chamber, and a plurality of gas sensors mounted within the sensing chamber.

The invention discloses a mobile robot positioning method based on multi-sensor combination, and belongs to the technical field of outdoor mobile robot positioning, and the method comprises the following steps: S1, sensor parameter calibration; s2, data acquisition and preprocessing; and S3, data fusion positioning. The method can adapt to various complex environments such as indoor and outdoor coexistence and multiple obstacles, continuously provides reliable and high-precision positioning information for the autonomous service robot, has the advantages of low cost, simple layout, good real-time performance, high reliability and the like, and can be widely applied to various autonomous robots.

A method for operating at least one autonomous robot (2), in particular an autonomous vegetation working robot, preferably an autonomous lawn robot, within an operating area, the robot comprising (i) at least one electrically driven tool, (ii) at least one electric motion drive for moving the robot, and (iii) at least one photoelectric device for converting energy from illuminating light, in particular sunlight, into electric energy for the tool and the motion drive, and (iv) at least one energy storage for storing electric energy charged by the photoelectric device and for sup-plying the tool and the motion drive with electric energy, the method comprising the steps of a) the robot (2) working in an autonomous working mode within at least one working area (15) within the operating area, b) determining an electric energy stored in the energy storage of the robot or a directly associated electric quantity such as electric capacity or electric charge of the energy storage, c) if the determined stored energy or the directly associated electric quantity of the energy storage is below a minimum operating threshold, the robot moves or is moved to a charging position (CP, WP) for recharging the energy storage using the photoelectric device, d) wherein the charging position (CP, WP) is chosen from one or more charging positions, where, at least for a minimum recharging time period, the illumination intensity of the illuminating light will be higher than a high illumination threshold and / or where the illuminating light will essentially not be shaded by objects in or around the operating area, in particular vegetation objects such as hedges, bushes or trees or built objects such as buildings, e) wherein the or each charging position (CP) is defined by a corresponding position marker (5) placed at the charging position, f) wherein the robot, when moving or being moved to the charging position, searches for the position marker (5) by means of a sensing device (25) and stops at the charging position within a certain distance from the position marker (5).

Some aspects include a method for operating an autonomous robot, including: capturing, with a first sensor disposed on the robot, data of an environment of the robot; generating, with the processor, a map of the environment based on at least the data of the environment; localizing, with the processor, the robot within the environment; capturing, with a second sensor disposed on the robot, data of a floor surface; determining, with the processor, a floor type of areas of the environment based on the data of the floor surface; and determining, with the processor, settings of the robot based on at least the floor type of the floor surface, wherein the settings comprise at least an elevation of each of at least one component of the robot from the floor surface.

A robotic delivery system includes a server, a dispatch module, and a robotic delivery vehicle. The dispatch module transmits a suggested item to the dispatch application being executable on a client device operated by a user; and receives, from the dispatch application, a selection of the suggested item from the user of the client device.

Robots, systems, methods, and computer program products for training and operating (semi-)autonomous robots to complete work objectives are described. A robot accesses a library of reusable work primitives from a catalog of libraries of reusable work primitives, each reusable work primitive corresponding to a respective basic sub-action that the robot is trained to autonomously perform. A work objective is analyzed to determine a sequence of reusable work primitives that complete the work objective, and the robot executes the sequence to complete the work objective. A robot can be deployed with access to an appropriate library of reusable work primitives, based on expectations for the robot. The robot is trained to perform reusable work primitives in multiple libraries, by generating control instructions which cause the robot to perform each reusable work primitive. Training is performed by real-world robots performing reusable work primitives, or simulated robot instances performing the reusable work primitives.

The present application relates to a kind of self-robot multi-task operation planning method and system driven by embodied cognitive large model, wherein the method comprises: S1, based on the real-time acquisition RGB diagram and depth diagram are encoded to obtain embodied visual representation;S2, obtain natural language instruction and and embodied visual representation are fused across modalities, obtain fusion feature, and based on fusion feature multi-task decomposition scheme;S3, based on multi-task decomposition scheme, utilize diffusion strategy to generate the continuous action trajectory of robot end effector;S4, obtain the second RGB image and second depth diagram after robot executes according to continuous action trajectory, and it is used as closed loop feedback signal, system is used to realize the above-mentioned method.Compared with prior art, the present application is based on embodied cognitive large model, and the depth of visual and language modalities is deeply fused, and combined with diffusion strategy action decision module, accurate action trajectory is predicted, and the multi-task autonomous planning and accurate execution capability of robot in dynamic complex environment is significantly improved.

Systems, methods, and computer program products for generating training data are described. Action data and context data are recorded for a robot body performing an action or task in an environment. The context data is augmented virtually to include variations from the recorded environment while the action data remains unchanged, and instances of training data are generated including the augmentations, to produce a large and varied training data set.