1097 results about "Industrial robotics" patented technology

The invention relates to the technical field of industrial automatic welding and discloses an intelligent control method and system for industrial robot welding, and the method comprises the steps that after a welding task is started, an initial welding process parameter model is called based on the current workpiece specification, and a first welding control instruction sequence is generated; according to the first welding control instruction sequence, a welding gun is controlled to execute welding operation, and multi-source sensing data flow in the welding process is synchronously collected; carrying out fusion feature extraction on the real-time sensing data stream to generate a current welding situation feature vector, when situation deviation exists between the current welding situation feature vector and the reference welding situation feature vector, obtaining a self-adaptive welding process parameter model through online incremental learning, and obtaining a self-adaptive welding process parameter model on the basis of the self-adaptive welding process parameter model. And generating an optimized welding control instruction sequence and issuing the optimized welding control instruction sequence to a robot controller and a welding gun so as to execute self-adaptive welding operation. The welding quality stability and the production efficiency can be improved, and the equipment maintenance cost is reduced.

The invention relates to the technical field of robot control, particularly discloses a path planning and dynamic obstacle avoidance control method for multi-task collaborative operation of an industrial robot, and aims to solve the problems of task scheduling conflict, dynamic obstacle response lag and low collaborative efficiency in a multi-robot system. The method comprises the following steps: constructing a task-resource joint scheduling model and generating initial task allocation; planning a conflict-free collaborative path based on an improved space-time A star algorithm; predicting a dynamic obstacle trajectory by using an LSTM network and generating a space-time envelope; constructing a second-order safety barrier function fusing task priorities; local obstacle avoidance re-planning is realized through rolling horizon model predictive control; and the global rescheduling is triggered when the task delay exceeds the limit or the deadlock risk occurs. According to the technical scheme, closed-loop linkage of global task collaboration and local dynamic obstacle avoidance is realized, and the system collaboration efficiency, the obstacle avoidance success rate and the operation robustness are remarkably improved.

The invention relates to the technical field of industrial robot control, and discloses a robot end trajectory optimization method and system based on industrial vision, and the method comprises the steps: obtaining a workpiece surface image in real time through a vision sensor, and obtaining deformation data through image enhancement and feature extraction; when the deformation quantity exceeds a threshold value, calculating a multi-axis coordination parameter by adopting an optimization algorithm to generate a trajectory correction instruction; integrating speed constraints by updating a control model, and determining a final trajectory by using Kalman filtering to fuse sensor feedback data; and the control parameters are iteratively adjusted in the deformation feedback loop, and a stable machining track is formed. The method can solve the problem that in the prior art, the robot tail end track precision is low.

The invention relates to the technical field of robot control, and discloses an industrial robot adaptive control method and system based on multi-modal sensor fusion, and the method comprises the steps: collecting multi-modal original data, and carrying out the time-space alignment; capturing space-time semantic association of visual textures, tactile pressure distribution and force sense fluctuation in the multi-modal data through a multi-head attention mechanism guided by a physical model, and performing space-time registration; a CNN-LSTM hybrid model is adopted to extract visual texture features and time sequence tactile features in the physical information enhanced multi-modal feature matrix; and carrying out dynamic weight distribution on the fusion feature vectors with physical consistency by utilizing a weight distribution model driven by element reinforcement learning to generate dynamic weighted fusion features. According to the method, the spatial positioning precision of the industrial robot in a precise assembly scene is greatly improved, the contact force control stability is greatly improved, and the control robustness in a complex operation scene is remarkably enhanced.

The invention provides a servo press control system and method based on a time-sensitive network, and relates to the technical field of intelligent manufacturing and industrial automation control, and the system comprises a time-sensitive network TSN which is used for providing global time synchronization and deterministic data transmission; a main controller of the multi-axis servo control unit is used for issuing a motion control instruction and a feeding and discharging control instruction through a TSN. A robot controller in the robot feeding and discharging unit is used for driving an industrial robot to execute plate taking and placing operation under the action of the feeding and discharging control instruction. The visual monitoring unit comprises an industrial camera accessed to the TSN and is used for collecting images and transmitting image data to the edge computing node through the TSN; the edge computing node is used for processing the image data to generate a compensation instruction and feeding back the compensation instruction through the TSN; and the upper monitoring system is used for subscribing real-time data published by the edge computing nodes, so that the control range of the servo pressure control system is expanded, and the convenience of data transmission and interaction is improved.

The invention discloses a data acquisition-based dynamic path planning method and system for an intelligent industrial robot with a body, and relates to the technical field of industrial robots. The method comprises the core steps of system calibration and parameter presetting before machining, multi-mode sensor dynamic acquisition, point cloud preprocessing, curvature sensing dynamic ICP registration, micro-inertia disturbance compensation, dynamic path planning, procedure adaptation adjustment, real-time feedback correction and the like. The point cloud quality is guaranteed through multi-sensor cooperation and dynamic switching, the registration effect is adjusted and optimized through curvature correlation parameters, dynamic compensation is achieved through a disturbance prediction model, the connection efficiency is improved through procedure parameter mapping, and accurate fitting of paths is guaranteed through graded feedback correction. The method is suitable for machining complex curved surfaces such as artificial joints, machining precision and stability can be remarkably improved, and the method has extremely high scene migration capacity, can provide reliable technical support for precision machining of similar complex curved surfaces and has remarkable popularization value.

The invention discloses an automatic driving carrying equipment scheduling system and method for an industrial robot. The system comprises a plurality of automatic driving carrying devices, a central dispatching device and corresponding communication modules. The system adopts a multi-agent reinforcement learning framework and combines a graph neural network processing environment topological structure to realize dynamic path planning and multi-device collaborative scheduling; integrating an energy consumption prediction mechanism based on a Kalman filter, and bringing energy consumption factors into a task allocation decision; meanwhile, a fault-tolerant management mechanism based on a distributed account book and federated learning is established, and fault detection and rapid recovery are achieved. The technical modules are deeply coupled, and a unified collaborative optimization framework is formed through reward function design, utility function optimization and fault probability calculation of multi-agent reinforcement learning. According to the method, the problems of poor dynamic environment adaptability, isolated decision making of each module, extensive energy consumption management and the like in the prior art are effectively solved, and the overall efficiency, energy efficiency and reliability of a scheduling system are remarkably improved.

The invention relates to the technical field of industrial robots, in particular to a dynamic simulation damping parameter solving method based on a gravity compensation mechanism, which comprises the following steps of: acquiring a joint basic parameter set of a to-be-tested industrial robot, and determining a test boundary condition according to the joint basic parameter set, and constructing a test task sequence based on the test boundary conditions. According to the method, the fusion weight of the static stiffness and the dynamic stiffness is intelligently adjusted according to the characteristics of the movement speed, the acceleration or the force control frequency band in a specific application task, and the limitation of a single parameter under complex and changeable working conditions is avoided. According to the parameter fusion mechanism aiming at the specific task characteristics, the generated joint comprehensive rigidity parameters can truly reflect the physical response of the robot in different operation stages, and the prediction error between the physical simulation environment and the actual physical operation is reduced; and the strict requirements of high-precision scenes such as semiconductor packaging and the like on micron-level positioning precision and force control stability are met.

The invention relates to an intelligent control method, system and device for an industrial robot and a medium. The method comprises the steps that high-dimensional coding is conducted on the environment and the body state of the robot to generate a unified hidden state vector; generating a task-driven original control instruction through a deep reinforcement learning network based on the hidden state vector, predicting the safety risk of the instruction through a safety network in a prospective manner, and outputting safety measurement and gradient data; when the risk is predicted to exist, constructing and solving an optimization model taking minimization intervention as a target, and calculating to obtain a safety correction value; fusing the correction value and the original instruction to generate a final safety control instruction and executing the final safety control instruction; and updating the security prediction network online by using the executed real security feedback. The method solves the problems that in existing deep learning control, the decision risk is unpredictable in physical deployment, and a traditional safety protection mechanism extensively interrupts tasks, and the intelligent decision-making ability of the robot is reserved to the maximum extent while absolute safety is guaranteed.

The invention discloses a real-time prospective trajectory planning method and device for an industrial robot and the robot, and the method comprises the steps: carrying out the discretization of a target trajectory, and constructing a corresponding speed constraint curve; dynamically adjusting the look-ahead window according to the running state of the controller; performing speed planning in a prospective window, generating a planning curve, comparing the planning curve with the speed constraint curve, and when the planning curve exceeds the limitation of the speed constraint curve, reducing the maximum speed and iterating the planning curve; generating a transition trajectory section meeting the high-order continuity of the speed curve at the joint in the target trajectory; when sudden stop or path reversal occurs, reconstructing a look-ahead window and generating a recovery section; and issuing each interpolated point obtained by planning to a servo system in real time for execution. According to the method, the risk of data hunger is reduced by dynamically adjusting the look-ahead window, meanwhile, the track continuity and the system stability are improved by supplementing the abnormal working condition recovery step, and the execution efficiency and precision of the robot are remarkably improved.

The invention discloses an intelligent grabbing system anomaly detection and processing method and device based on 3D vision, and belongs to the technical field of industrial robot 3D vision guide grabbing. The method comprises the steps of collecting 3D point cloud data of stacked flexible plates; carrying out point cloud preprocessing and template matching, and determining a plate pose and a grabbing point; generating a cushion block detection area and a deformation detection area based on the grabbed point pose; counting the number of point clouds in the detection area, comparing the number with a threshold value, and judging whether cushion block residues or plate excessive deformation exists or not; if yes, triggering processing, and if not, outputting a grabbing instruction. The device comprises a truss manipulator, an end effector, a 3D visual sensor, a visual processing unit and a robot control unit. According to the method, the detection area is dynamically generated, point cloud threshold judgment is fused, self-adaptive detection and processing of excessive deformation of the cushion block and the plate are achieved, the robustness, safety and automation level of a grabbing system are remarkably improved, and the method is suitable for automatic unstacking operation of stacked flexible plates.

The embodiment of the invention provides an auxiliary calibration method of a two-dimensional vision camera, and the method comprises the steps: starting a conveying belt, recording an initial value of a motor encoder, and guaranteeing the accuracy of data recording; the calibration plate is moved to the preset position associated with the industrial robot, and the workbench plane calibration is performed according to the numerical value change condition of the motor encoder, so that the installation inclination error of the workbench can be effectively compensated, and the calibration precision is improved; the industrial robot is controlled to move to the position above a nine-point grid in sequence at the following position of the calibration plate, physical coordinates are recorded, a mapping relation is established in combination with pixel coordinates shot by a camera, a conversion matrix is determined, automatic conversion between a robot coordinate system and a camera coordinate system is achieved, and calibration efficiency and consistency are remarkably improved; the passive electromagnetic calibration needle is matched with the calibration plate with the built-in two-dimensional electrified wire grid, and high-precision positioning is realized by utilizing an electromagnetic induction principle, so that the complexity and the cost of the system are reduced, and meanwhile, the reliability and the repeatability of the calibration process are ensured.

The invention discloses an AI-driven industrial robot intelligent assembly collaborative scheduling method and system, and the method comprises the steps: collecting assembly operation data, and constructing a semantic structure which comprises a process, a robot and a space region; extracting time sequence data, inputting the time sequence data into the prediction model, calculating offset and mapping a shell state; constructing an origami surface structure, analyzing space and time conflicts, and determining conflict positions; rearranging execution space and time windows, generating constraint mapping, and constructing a regular space-time layout; constructing a two-dimensional scheduling texture block, executing a reconstruction operation, and generating a scheduling texture sequence; and generating a scheduling scheme, completing constraint verification, and issuing the scheduling scheme to a robot control system for execution. According to the method, by introducing a beat prediction and state driving cooperative scheduling mechanism based on assembly semantic data, advanced perception and stable cooperative control of beat fluctuation, space conflicts and scheduling disturbance in the assembly process of multiple industrial robots are achieved.

The invention discloses a high-flexibility multifunctional cable and a manufacturing process and application thereof, and relates to the technical field of communication cables.The high-flexibility multifunctional cable comprises at least one conductor unit, and each conductor unit comprises a conductive wire core formed by twisting a plurality of metal conductors and a first inner insulating layer wrapping the exterior of the conductive wire core; the functional unit and the conductor unit are arranged in a bundling mode, the functional unit comprises a functional core body, a second inner insulating layer and a lapped inner shielding layer, the functional core body is wrapped with the second inner insulating layer, and after the conductor unit and the functional unit are combined, the conductor unit and the functional unit are sequentially wrapped with an outer shielding layer, a buffer layer, an armor layer and an outer sleeve layer from inside to outside. According to the invention, application scenes in main connecting pieces such as electric power communication composite power supply / transmission umbilical cords, aerospace cabin-out system umbilical cords, industrial robot automatic mechanical arms and the like can be met.

The utility model relates to a pillow spring assembling and disassembling clamp which comprises an electromagnetic iron core, a coil winding, a pneumatic tool, an ejector pin, a pen-shaped air cylinder, an adjusting gasket, a supporting seat and an installing seat, the electromagnetic iron core is fixedly arranged on the supporting seat, a combination mode of the electromagnetic iron core and the pen-shaped air cylinder is adopted, and the magnetic pole face of the electromagnetic iron core adopts two sections of cambered surfaces and a plane. The two sections of cambered surfaces are used for magnetically attracting bearing springs and damping springs of different sizes respectively, the plane is used for magnetically attracting a wedge or other parts, and the height is set to be the height of the two rings of springs so that the industrial robot can work in the side frame conveniently; the cylinder mounted behind the electromagnet adopts the design of a single-pen cylinder and a round head ejector pin, so that the problem of clamping of multiple cylinders is avoided, the ejector pin round head structure effectively realizes clamping and releasing of an inner spring and an outer spring, the mechanical failure rate is reduced, and automatic disassembly and assembly operation of different types of bogie pillow springs, wedges and other parts is facilitated.

The invention discloses an automatic product grabbing system and method, and relates to the technical field of industrial robots, in the system, an equipment body is provided with a grabbing station and a placing station, an inbound positioning assembly outputs an AGV tool vehicle in-place signal, and a jacking positioning assembly establishes a preset grabbing height; the visual positioning assembly collects an image output pose compensation amount, the bar code recognition assembly obtains a bar code and interacts with the manufacturing execution system to obtain a verification result, the controller issues the compensation amount to the industrial robot control end when verification is allowed, and an end effector is driven to complete positioning, locking and unlocking of grabbing locked by a tool vehicle clamp; the tray transferring and positioning mechanism completes transferring and positioning of the shipping tray, and the unlocking and locking mechanism unlocks / locks the locking block handle and conducts visual review in the placing process. By means of the technical scheme, the grabbing and placing alignment precision and the mistake proofing capacity are improved, the risks of collision, falling and locking omission are reduced, and the takt time stability and traceability are improved.

The invention discloses an air spring air chamber screw embedding device which comprises a feeding mechanism arranged on one side of a preheating mechanism, and a first transferring mechanism is arranged between the feeding mechanism and the preheating mechanism; the second transferring mechanism is connected to the machine table, and the feeding end of the second transferring mechanism is connected with a coding mechanism. The industrial robot A is arranged on one side of the injection molding machine, and the execution end of the industrial robot A is connected with a second manipulator; the third transfer mechanism is arranged on one side of the second transfer mechanism, an industrial robot B and a detection mechanism are arranged on the two sides of the second transfer mechanism, and the execution end of the industrial robot B is connected with a third manipulator; a fourth mechanical arm is connected to the fourth transferring mechanism, and a weighing mechanism and a discharging mechanism are arranged on the portion, below the fourth transferring mechanism, of the machine table. Compared with the prior art, feeding, preheating, embedding, coding, detecting, weighing and discharging are integrated, time and labor are greatly saved, the working efficiency is improved, and meanwhile the product quality is guaranteed.

A method of operating a computer-controlled industrial robotic arm to sharpen a knife can include performing a profilometer scan of the knife and determining the scan is incomplete or missing data. Rather than re-scanning the knife, a template having a similar geometry to completed sections of the scan can be selected from a library of template knife profiles. The templates in the library can be generated from previously completed scans of other knives. A completed proxy knife geometry can be created by substituting the template geometry for missing scan regions or in place of the incomplete scan. The completed proxy knife geometry can then be used for controlling a robotic arm to perform a sharpening operation upon the knife.

The invention relates to the technical field of industrial robot precision machining, in particular to an intelligent monitoring and process parameter self-adaption regulation and control method and system which are suitable for metal plate precision cutting, composite material special-shaped part cutting, pipe three-dimensional contour cutting and other scenes. Real-time state monitoring, cut quality online evaluation, process parameter dynamic optimization and equipment health early warning in the cutting process can be achieved, and the problems that in traditional laser cutting, cut quality consistency is poor, the parameter debugging period is long, and the material utilization rate is low are solved.

The invention relates to the field of precision manufacturing of aero-engine blades, and particularly discloses an aero-engine blade grinding and polishing removal amount control method which comprises the steps that an industrial robot carries 3D visual equipment to conduct multi-angle scanning on a blade pasted with mark points, the number of public mark points between adjacent scanning poses is forcibly restrained to be larger than or equal to 3, and the number of the mark points is smaller than or equal to 3; a rigid body transformation matrix is solved based on space coordinate transformation, and high-precision point cloud automatic splicing is achieved; blade point clouds before and after grinding and polishing are reconstructed through Delaunay triangulation with consistent parameters, and a triangular mesh continuous curved surface is generated; calculating the minimum space distance from each vertex of the target curved surface to the source curved surface as a removal amount error epsilon, and positioning an out-of-tolerance region based on epsilon distribution; and generating a compensatory processing track by adopting an equal residual height method. The problem of point cloud registration caused by rare geometric features of the blade is solved, machining decision is driven through curved surface space errors, closed-loop accurate control over the removal amount is achieved, and the consistency of the grinding and polishing quality of the blade is improved.

The invention discloses a welding robot work monitoring method and system based on machine vision. According to the scheme, a monitoring platform composed of an industrial robot, a welding head, a machine vision sensor, an industrial personal computer and a controller is constructed; on the basis, a unified coordinate system is established through camera internal reference calibration, hand-eye calibration and tool center point calibration; welding images, robot poses and process parameters are collected, and welding seam features are extracted through preprocessing and dynamic region of interest locking; performing three-dimensional reconstruction and track fitting by combining the depth information to generate a welding seam reference track and a welding gun theoretical attitude; and finally, the working state is judged by fusing the deviation parameters, the track and the technological parameters, time delay compensation, online deviation correction and abnormal alarm are carried out, and the scheme can improve the welding seam monitoring precision, the tracking stability and the welding quality consistency under the strong interference working condition.

The invention relates to the technical field of industrial robot path planning, and discloses a path planning obstacle avoidance system of an industrial robot. According to the system, a space grid model containing obstacles is constructed through an environment modeling module. The obstacle processing module identifies an obstacle profile and generates an affected area thereof. And the feasible space analysis module calculates and divides a passable area of the robot according to the result. The trajectory initialization module generates an initial reference trajectory in a designated area and calculates a quantized avoidance priority parameter for each waypoint. And the path optimization and output module identifies and adjusts high-risk key points in the trajectory according to the priority parameters, performs iterative optimization in a feasible space, and finally outputs a safe obstacle avoidance path. According to the technology, differentiated evaluation and priority-driven optimization are carried out on the risk of the path points, so that the safety and the overall efficiency of path planning in a complex environment are improved.

The invention provides a three-dimensional steel structure full-coverage path planning inspection method based on topology reconstruction, and relates to the technical field of industrial robots, and the method comprises the steps: obtaining a model file of a to-be-inspected three-dimensional steel structure; constructing an original vertex set and an original edge set according to the model file; performing adaptive clustering on each original vertex in the original vertex set through a density-based spatial clustering algorithm to obtain a cluster; taking the geometric center of the original vertex in each cluster as a topological node, and establishing a mapping relation between the original vertex and the topological node; carrying out topology reconstruction on the original edge set, and constructing an undirected weighted graph; stitching non-connected sub-graphs in the undirected weighted graph through a K-dimensional tree and a minimum spanning tree algorithm to obtain a fully connected graph; generating a full-coverage continuous path sequence through a Chinese postman algorithm; the full-coverage continuous path sequence is converted into a three-dimensional space coordinate sequence, and the three-dimensional steel structure to be inspected is inspected according to the three-dimensional space coordinate sequence.

This invention discloses a power cord anti-pull mechanism and an industrial six-axis robot, belonging to the field of industrial robot technology. The power cord anti-pull mechanism is applied to equipment requiring electrical power for operation. It includes a power cord, a power supply box, and an anti-pull device. One end of the power cord is equipped with a power plug, and the other end is used to connect to the equipment requiring electrical power. A power socket is provided on one side of the power supply box, into which the power plug can be inserted to connect to the power supply. The anti-pull device is disposed on the power supply box and includes a fastening device and a power device. The fastening device is used to secure the power plug after it is inserted into the power socket, preventing pulling between the power cord and the power socket. The power device provides power to the fastening device to secure the power plug. The power cord anti-pull mechanism can prevent problems such as poor contact or disconnection between the power cord and the power supply box during the operation of the industrial six-axis robot.

The invention relates to an ultrasonic detection and damage evaluation method and system for a curved-surface composite component, and the method comprises the steps: collecting three-dimensional point cloud data of a surface damage region of the curved-surface composite component, and constructing a detection path point sequence; the industrial robot and the ultrasonic phased array are controlled to move and scan synchronously according to time, and ultrasonic A scanning signals and robot tail end pose information are collected; three-dimensional space coordinates of all reflection points in the curved surface composite material component are calculated in combination with pose derivation, and an initial three-dimensional damage point cloud is constructed; setting a segmentation threshold to construct an initial segmentation damage point cloud; a traditional DBSCAN clustering algorithm is improved for clustering segmentation, and point clouds of all independent damage areas are obtained; and voxelizing the point cloud of the independent damage area into a three-dimensional voxel model to realize quantitative assessment of the damage. According to the method, automatic and high-precision three-dimensional damage detection and evaluation of the complex curved surface component are realized, and the problems that a traditional ultrasonic detection method depends on manual operation and is poor in adaptability, the detection result is not visual and the like are effectively solved.

The invention provides a training method of a visual sense and tactile sense language action large model, the visual sense and tactile sense language action large model and a robot, and relates to the technical field of industrial robots. The method comprises the following steps: training a bridging module between a task planning system and an action execution system; training the task planning system; and finally, performing end-to-end supervision fine tuning on the visual and tactile language action large model. Task planning and an action execution function are decoupled, and a tactile encoder and a three-dimensional point cloud encoder are introduced in stages, so that the model can fuse tactile information in a planning stage to understand physical interaction, and point cloud data can be utilized to realize accurate positioning in an execution stage; by adopting the three-stage training method, the training stability and convergence efficiency of a multi-mode and multi-task model are effectively guaranteed, the operation precision, success rate and environmental adaptability of a robot in industrial assembly tasks are fundamentally improved, and the practical application of a visual language action model in industrial flexible production is promoted.

The invention discloses an industrial robot motion pose error calibration method based on constraint solution. The industrial robot motion pose error calibration method comprises the steps that S1, joint angles, tail end coordinates and temperature data of multiple measurement points are collected; s2, correcting the kinematic model by using the thermal elongation and constructing a parameter identification model; s3, constructing a mixed integer nonlinear constraint pool according to mechanical limitation; s4, introducing a semi-definite programming slack variable to convert the non-convex constraint into a linear matrix inequality; s5, combining a primal-dual interior point method and an alternating direction multiplier method to iteratively solve the optimal kinematics parameter correction; s6, introducing a square root information filtering improvement mechanism to correct a state vector square root factor, and calculating an updated nominal kinematics parameter; and S7, based on the condition number of the error Jacobian matrix, judging observability and repeatedly calibrating. According to the method, the problem of precision reduction caused by thermal errors is solved, and the identification precision, the numerical stability and the self-adaptive calibration capability of the kinematics parameters of the robot are improved.

The invention discloses a robot fine operation control system based on a diffusion model and 3D vision, particularly relates to the technical field of industrial robot fine operation control, and aims to solve the problem that when an existing industrial robot executes fine operations such as bottle cap screwing and connector plugging in and out in a complex three-dimensional operation scene, the operation efficiency is high. The problems that three-dimensional perception is unstable, the action resolution is fixed, space and contact constraints are difficult to meet, and the task success rate and safety are insufficient are solved. The method comprises the following steps: constructing a voxel feature field updated along with time under a self-body coordinate system, generating interaction complexity and a physical constraint field based on task coding, and linking adaptive resolution and action granularity scheduling, diffusion trajectory generation and closed-loop execution correction to uniformly map environmental geometry, contact conditions and task stages into tail end trajectory control, so as to realize real-time tracking control. Therefore, fine operations such as bottle cap screwing, connector inserting and pulling and flexible object folding can still be stably completed under complex shielding and contact constraint, and the effects of safety and task success rate are both considered.

The invention discloses an industrial inspection abnormity identification method for an inspection robot based on a multi-modal large model, which creatively constructs an intelligent identification architecture of'real-time initial inspection of a visual small model + secondary recheck of the multi-modal large model ', takes the strong basic cognitive ability of the multi-modal base large model as a foundation stone, and combines with an efficient parameter LoRA fine tuning technology to identify the industrial inspection abnormity of the inspection robot. Through industrial knowledge graph injection and expert rule guidance, Few-Shot learning is carried out only by using a small amount of annotation data of a current scene, and a special multi-modal large model which deeply understands a specific industrial environment and has a multi-dimensional reasoning capability can be rapidly obtained, so that the difficulty and period of high-quality data collection and model training are greatly reduced, and the working efficiency is improved. And through a double-layer processing mechanism and an uncertainty quantification technology, while the real-time response capability is ensured, the anomaly identification accuracy and generalization are remarkably improved, and the method is suitable for various industrial robot and automatic production line monitoring scenes with extremely high reliability requirements.

The invention relates to the technical field of industrial robot control, and discloses a battery pack flexible assembly line and deviation cooperative control method and system based on machine vision. The method comprises the following steps: constructing a digital twin model synchronous with a physical production line, and defining a key feature point network representing assembly precision; for the current assembly station, generating a feedforward pre-adjustment instruction through a preset cooperative control strategy; acquiring real-time assembly data by utilizing a plurality of sensors deployed at a current station, and generating a real-time feedback fine tuning instruction; the collaborative decision module fuses the feedforward and feedback instructions to generate a final execution instruction, and drives the assembly robot to complete high-precision assembly; and finally, the assembly result data is transmitted back to the digital twin model for dynamically optimizing the cooperative control strategy. The problems of deviation accumulation, passive compensation and low production line remodeling efficiency in multi-station assembly are solved, and production line-level deviation active collaborative management and flexible production are achieved.