A multi-degree-of-freedom control method, system, product and medium for a floor-track robot

Abstract

A multi-degree-of-freedom control method, system, product and medium for a ground rail robot, which relates to the field of control or regulation systems. The method includes: extracting the characteristic quantities in the handling object task data to obtain the weight parameter, material parameter and position parameter of the handling object, matching the track operation data of the ground rail in the preset track database according to the weight parameter, material parameter and position parameter, extracting the target points in the track path of the ground rail, calculating the handling path data of the robot on the ground rail based on the handling parameters of the target points, sending the handling path data to the robot, after the robot executes the handling path data, measuring the real-time force parameters and real-time position parameters of the robot at the picking point, transition point and placing point, inputting the real-time force parameters and real-time position parameters into the preset evaluation model to obtain the track correction parameter and position correction parameter, and sending the correction instruction to the ground rail robot. Implementing this method can improve the handling accuracy of the ground rail robot.

Description

Technical Field

[0001] This application relates to the field of general control or regulation systems, and particularly to a multi-degree-of-freedom control method, system, product and medium for a ground-rail robot. Background Art

[0002] With the continuous improvement of industrial automation, ground-rail robot systems are widely used in intelligent factories to complete material handling tasks. The ground-rail robot system realizes precise material handling and positioning placement through robots running on fixed tracks, improving production efficiency while reducing labor costs.

[0003] In the prior art, ground-rail robot systems usually control with preset track movement paths. Before the system is put into use, staff will plan the movement trajectory of the robot in advance according to the factory layout and production requirements, and solidify the corresponding control parameters in the control system. The ground-rail robot executes the handling task according to the preset track movement path, and performs positioning and path tracking through position sensors.

[0004] However, in the actual production process, due to the differences in the weight and material of the handled materials, and the vibration and deformation generated during the long-term operation of the ground rail, the preset track movement path is difficult to adapt to the dynamically changing working conditions. This fixed control method is likely to cause the handled materials to deviate or shake, affecting the handling accuracy and even possibly causing material damage. Summary of the Invention

[0005] This application provides a multi-degree-of-freedom control method, system, product and medium for a ground-rail robot, which is used to improve the handling accuracy of the ground-rail robot.

[0006] In a first aspect, the present application provides a multi-degree-of-freedom control method for a gantry robot, which is applied to a multi-degree-of-freedom control system of a gantry robot. The multi-degree-of-freedom control system of the gantry robot includes a gantry robot and a gantry. The method includes: obtaining handling object task data, and extracting characteristic quantities from the handling object task data to obtain the weight parameter, material parameter, and position parameter of the handling object; matching the track operation data of the gantry according to the weight parameter, the material parameter, and the position parameter in a preset track database, where the track operation data includes track vibration parameters, speed parameters, and displacement parameters; extracting target points in the track path of the gantry, and calculating the handling path data of the gantry robot on the gantry based on the handling parameters of the target points, where the target points include a pickup point, a transition point, and a placement point; sending the handling path data to the gantry robot, and after the gantry robot executes the handling path data, measuring the real-time force parameters and real-time position parameters of the gantry robot at the pickup point, the transition point, and the placement point; inputting the real-time force parameters and the real-time position parameters into a preset evaluation model to obtain a track correction parameter and a position correction parameter; and sending a correction instruction including the track correction parameter and the position correction parameter to the gantry robot.

[0007] By adopting the above technical solution, first, the weight, material, and position parameters of the handling object are obtained. Then, the target points of the gantry are extracted, and the handling path data is calculated in combination with relevant parameters and sent to the robot. After the robot executes, its force and position parameters are collected and input into the preset evaluation model to obtain the correction parameters and then sent back to the robot. The handling object parameters provide a basis for matching the track operation data. The calculated handling path data guides the movement of the robot. The correction parameters obtained based on the actual state of the robot can adjust its movement in a timely manner, effectively improving the handling accuracy and stability, and ensuring the high-quality completion of the handling task.

[0008] Combined with some embodiments of the first aspect, in some embodiments, the step of extracting target points in the track path of the gantry and calculating the handling path data of the gantry robot on the gantry based on the handling parameters of the target points specifically includes: obtaining the initial positions of the pickup point, the transition point, and the placement point; adjusting the initial positions of the pickup point, the transition point, and the placement point according to the track vibration parameters to obtain adjusted positions; calculating the movement trajectories of the gantry robot between the points according to the displacement parameters; calculating the speed change sequence of the gantry robot during the movement process according to the speed parameters; and converting the adjusted positions, the movement trajectories, and the speed change sequence into the control instruction format of the gantry robot to generate the handling path data.

[0009] By adopting the above technical solution, when calculating the handling path data, first obtain the initial position of the target point, adjust it according to the track vibration parameters, then calculate the motion trajectory and the speed change sequence respectively using the displacement and speed parameters, and finally generate the data after converting the format. The adjustment of the track vibration parameters makes the position of the target point more reasonable, avoiding deviation caused by vibration. The displacement parameter ensures the accuracy of the motion trajectory of the robot between each point, and the speed parameter ensures that the motion speed meets the requirements. The generated handling path data can better adapt to the actual situation of the track, making the robot move more precisely and efficiently, and improving the overall performance of the handling operation.

[0010] Combined with some embodiments of the first aspect, in some embodiments, before the step of inputting the real-time force parameter and the real-time position parameter into a preset evaluation model to obtain an orbit correction parameter and a position correction parameter, the method further includes: collecting the historical force parameter of the ground track during multiple operations and the historical position parameter of the ground track robot; extracting the historical force parameter to obtain a force feature, and extracting the historical position parameter to obtain a position feature; training a neural network model based on the force feature and the position feature to obtain the preset evaluation model, where the neural network model includes a first sub-network for generating an orbit correction parameter and a second sub-network for generating a position correction parameter.

[0011] By adopting the above technical solution, the historical force and position parameters of the ground track and the robot are collected, and after extracting the features, a neural network model is trained to obtain a preset evaluation model. The historical parameters provide a rich data basis for model training. By learning the rules and features in these data, the neural network model can accurately analyze the motion state of the robot and generate appropriate correction parameters. The first sub-network and the second sub-network of the model are optimized for the orbit and the position respectively, making the obtained correction parameters more targeted and effective, enhancing the correction ability of the robot's motion, and ensuring the stability and accuracy of the handling process.

[0012] Combined with some embodiments of the first aspect, in some embodiments, after the step of sending the attitude adjustment instruction including the attitude adjustment amount and the execution speed to the ground track robot, the method further includes: obtaining the real-time image data of the ground track robot, and identifying the attitude parameter of the handled object from the real-time image data; when the attitude parameter exceeds a preset attitude threshold, calculating the attitude adjustment amount according to the orbit correction parameter and the position correction parameter; determining the execution speed of the attitude adjustment based on the weight parameter and the material parameter; and sending the attitude adjustment instruction including the attitude adjustment amount and the execution speed to the ground track robot.

[0013] By adopting the above technical solution, the attitude parameters of the transported object are obtained based on real-time image recognition. When the attitude exceeds the threshold, the attitude adjustment amount is calculated and the execution speed is determined according to the weight and material parameters. The track correction parameter reflects the track deformation state, and the position correction parameter reflects the position deviation of the robot. The two work together to generate the optimal attitude adjustment amount. The execution speed is matched with the characteristics of the transported object. Heavy objects use low speed to ensure stability, and light objects use high speed to improve efficiency. The precise execution of the attitude adjustment instruction keeps the transported object within a reasonable attitude range at all times, avoiding the damage risk caused by deviation and shaking. At the same time, the speed control considering the material characteristics ensures the safety and reliability of the adjustment process, achieving high-quality control of the handling process.

[0014] Combined with some embodiments of the first aspect, in some embodiments, after the step of sending the attitude adjustment instruction including the attitude adjustment amount and the execution speed to the ground rail robot, the method further includes: when the attitude adjustment instruction starts to be executed, collecting the real-time force parameters of the ground rail; when the real-time force parameter is greater than a preset safety threshold, sending a pause instruction to the ground rail robot to make the ground rail robot pause the attitude adjustment; when the real-time force parameter returns to within the preset safety threshold range, reducing the execution speed to a preset execution speed threshold.

[0015] By adopting the above technical solution, the real-time force parameters of the ground rail are monitored during the execution of the attitude adjustment. When the safety threshold is exceeded, the adjustment is immediately paused and waiting for recovery. The force parameter reflects the force state of the entire system, and pausing in time can prevent excessive stress from damaging the ground rail and the robot. When the force returns to normal, the adjustment is continued in the speed reduction execution mode, and reducing the speed effectively suppresses the impact and vibration. The pause and speed reduction mechanisms form a complete safety protection system, intelligently adjusting the execution state to avoid the occurrence of dangerous working conditions.

[0016] Combined with some embodiments of the first aspect, in some embodiments, after the step of inputting the real-time force parameter and the real-time position parameter into a preset evaluation model to obtain the track correction parameter and the position correction parameter, the method further includes: determining the track vibration threshold based on the weight parameter and the material parameter; during the operation of the ground rail robot, collecting the real-time track vibration parameters of each track segment; comparing the real-time track vibration parameter with the maximum allowable track vibration threshold, marking the track segments that exceed the maximum allowable track vibration threshold to obtain the marked track segments; sending the positions of the marked track segments to the target client.

[0017] By adopting the above technical solution, the vibration threshold of the track is determined according to the characteristics of the carried object, the vibration parameters of the track section are collected and compared with the threshold to mark the over-limit section. The vibration threshold is dynamically adjusted according to the material characteristics, and the weight and material parameters determine the maximum allowable vibration amplitude. The vibration state of each track section is monitored in real time, and the position information of the over-limit section is sent to the client in a timely manner. The track section marking mechanism realizes the accurate positioning of vibration anomalies, the dynamic threshold ensures the matching of the monitoring standard with the handling task, and the timely information feedback supports preventive maintenance, constructing a track monitoring system for active warning and fault location.

[0018] Combined with some embodiments of the first aspect, in some embodiments, after the step of inputting the real-time force parameter and the real-time position parameter into a preset evaluation model to obtain a track correction parameter and a position correction parameter, the method further includes: processing the real-time image data by using an AR-assisted visual recognition technology to obtain environmental feature parameters; adjusting the track correction parameter and the position correction parameter according to the environmental feature parameters.

[0019] By adopting the above technical solution, an AR-assisted visual technology is used to process real-time images to obtain environmental features, and the track and position correction parameters are adjusted accordingly. The AR technology superimposes virtual information on the real scene, enhancing the environmental perception ability and accurately capturing spatial geometric features, lighting conditions, and occlusion conditions. The environmental features directly affect the calculation of the correction parameters. The geometric features are used for track plane correction, the lighting conditions guide attitude compensation, and the occlusion information optimizes position adjustment. The parameter adjustment mechanism based on environmental features establishes a dynamic association between the correction amount and the actual working conditions, improves the adaptability of the correction strategy to environmental changes, and realizes more accurate track motion control.

[0020] In a second aspect, an embodiment of the present application provides a multi-degree-of-freedom control system for a ground track robot. The multi-degree-of-freedom control system for the ground track robot includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, and the computer program code includes computer instructions. The one or more processors call the computer instructions to enable the multi-degree-of-freedom control system for the ground track robot to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product containing instructions. When the above computer program product runs on the multi-degree-of-freedom control system for a ground track robot, it enables the multi-degree-of-freedom control system for the ground track robot to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0022] Fourthly, an embodiment of the present application provides a computer-readable storage medium, including instructions, which, when running on the multi-degree-of-freedom control system of the gantry robot, cause the multi-degree-of-freedom control system of the gantry robot to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0023] It can be understood that the multi-degree-of-freedom control system of the gantry robot provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiments of the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be elaborated here.

[0024] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0025] 1. In the present application, by first obtaining the weight, material, and position parameters of the carried object, then extracting the gantry target points and calculating the handling path data in combination with relevant parameters and sending it to the robot. After the robot executes, it collects its force and position parameters, inputs them into a preset evaluation model to obtain correction parameters and then sends them back to the robot. The parameters of the carried object provide a basis for matching the track operation data. The calculated handling path data guides the movement of the robot. The correction parameters obtained based on the actual state of the robot can adjust its movement in a timely manner, effectively improving the handling accuracy and stability, and ensuring the high-quality completion of the handling task.

[0026] 2. In the present application, by obtaining the attitude parameters of the carried object based on real-time image recognition, when the attitude exceeds the threshold, calculating the attitude adjustment amount and determining the execution speed according to the weight and material parameters. The track correction parameters reflect the track deformation state, and the position correction parameters reflect the position deviation of the robot. The two work together to generate the optimal attitude adjustment amount. The execution speed matches the characteristics of the carried object. Heavy objects use low speed to ensure stability, and light objects use high speed to improve efficiency. The precise execution of the attitude adjustment instruction keeps the carried object within a reasonable attitude range at all times, avoiding the damage risk caused by deviation and shaking. At the same time, the speed control considering the material characteristics ensures the safety and reliability of the adjustment process, realizing the high-quality control of the handling process.

[0027] 3. In the present application, by real-time monitoring the force parameters of the gantry when performing attitude adjustment, if it exceeds the safety threshold, immediately pause the adjustment and wait for recovery. The force parameters reflect the force state of the entire system. Pausing in a timely manner can prevent excessive stress from damaging the gantry and the robot. When the force returns to normal, continue to complete the adjustment in a reduced-speed execution mode. Reducing the speed effectively suppresses impact and vibration. The pause and reduced-speed mechanisms form a complete safety protection system, intelligently adjusting the execution state to avoid the occurrence of dangerous working conditions. Description of the Drawings

[0028] Figure 1It is a schematic flowchart of the multi-degree-of-freedom control method for the ground rail robot in the embodiments of the present application;

[0029] Figure 2 It is another schematic flowchart of the multi-degree-of-freedom control method for the ground rail robot in the embodiments of the present application;

[0030] Figure 3 It is a schematic structural diagram of an entity device of the multi-degree-of-freedom control system for the ground rail robot in the embodiments of the present application. Detailed implementation manners

[0031] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application, the singular forms "a", "an", "the above", "the", and "this" are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations including one or more of the listed items.

[0032] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as implying or indicating relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0033] For ease of understanding, the application scenarios of the embodiments of the present application are introduced below.

[0034] In the general assembly workshop of an automobile manufacturing plant, various automotive parts with different weights need to be transported from the material area to the assembly line. Multiple ground rails are installed in the factory, and the ground rail robots are responsible for transporting different materials such as engine assemblies (weighing 300 kg), door assemblies (80 kg), and instrument panel assemblies (50 kg). Due to the large difference in material weight, diverse materials, and uneven wear and local deformation of the ground rails after 3 years of operation, material shaking and position deviation often occur during the transportation process. Especially when transporting heavy engines, the position deviation caused by ground rail vibration reaches ±15 mm, seriously affecting the assembly accuracy. At the same time, light materials such as doors will also vibrate violently due to the uneven track surface during high-speed transportation, increasing the risk of surface scratches, and the material loss caused thereby exceeds 500,000 yuan per month.

[0035] A factory used the traditional fixed parameter control solution to deal with the above problems: a one-time calibration was performed when the ground rail was installed, the rail parameters and motion trajectory were solidified in the control system, and the same running speed (1.2m / s) and acceleration (0.5m / s²) were used for all materials. When carrying a 300kg engine, the robot strictly followed the preset trajectory, but because the deformation of the track was not taken into account, the engine shifted at the transition point. Workers had to manually adjust the engine position, which increased the assembly time by 30%. When carrying a 50kg instrument panel, the same low-speed running parameters were used, which affected the production rhythm. What's more serious is that when the track was deformed by 2mm, the control system could not be adjusted in time, causing the material to collide with the surrounding equipment, and 3 safety accidents occurred within a month.

[0036] After the factory adopts this application plan, it first identifies the characteristics of different materials through task data. When carrying a 300kg engine, the system automatically matches the track operation parameters, reduces the speed to 0.8m / s, and dynamically adjusts the operation trajectory according to the real-time vibration value of the track (1.5mm). When the force sensor detects an increase in force (800N) at the transition point, the system immediately calculates the correction amount and adjusts the robot posture to allow the engine to pass smoothly. For the 50kg instrument panel, considering its lighter weight, the system increases the speed to 1.5m / s. At the same time, the AR vision system monitors the surrounding environment in real time, and automatically optimizes the path planning when it identifies obstacles 1.2m away. During the entire handling process, the system continuously collects track segment data and immediately notifies maintenance personnel when the vibration of a certain section exceeds 2mm. Three months after implementation, material losses were reduced by 90% and assembly efficiency increased by 35%.

[0037] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above scenario. Figure 1 , which is a flow chart of a multi-degree-of-freedom control method for a rail robot in an embodiment of the present application.

[0038] S101, acquiring task data of a transported object, and extracting feature quantities in the task data of the transported object to obtain weight parameters, material parameters, and position parameters of the transported object.

[0039] Among them, the task data of the transported object represents the basic information and transport requirements of the object to be transported, including the basic attribute data of the object and the transport task parameters. The weight parameter refers to the mass of the transported object in kilograms. The material parameter is used to represent the material properties of the transported object, including hardness, elastic coefficient, etc. The position parameter represents the coordinate position and posture angle of the transported object in three-dimensional space. The feature quantity refers to the key numerical features extracted from the task data.

[0040] After receiving the handling task instruction, it is necessary to first obtain the basic data related to the task. Specifically, the system obtains the task data of the handled object through sensor acquisition or database query, and then uses a feature extraction algorithm to process the original data to extract three key parameters: weight, material, and position. The weight parameter is directly measured by a weighing sensor, the material parameter is obtained by matching a preset material database, and the position parameter is measured by a vision positioning system.

[0041] In some embodiments, parameter acquisition and feature extraction can be achieved in the following ways: Optionally, first collect multi-angle images of the handled object through an industrial camera, use a deep learning algorithm to identify the type of the handled object, query the preset database to obtain the standard weight and material parameters, and calculate the position parameter through a vision positioning algorithm; Optionally, use a weighing sensor to measure the weight in real time, measure the contact force through a force sensor to obtain the material characteristics, and use a lidar scanner to obtain three-dimensional position information. It can be understood that other sensing methods and algorithms can also be used to obtain the parameters of the handled object, which are not limited here.

[0042] S102. Match the track operation data of the ground rail in the preset track database according to the weight parameter, the material parameter, and the position parameter. The track operation data includes track vibration parameters, speed parameters, and displacement parameters.

[0043] Among them, the preset track database is a structured data set storing the operation parameters of the ground rail. The track operation data represents the dynamic characteristic parameters of the ground rail under different working conditions. The track vibration parameter is used to represent the vibration characteristics of the track, including amplitude and frequency. The speed parameter refers to the motion speed limit of the ground rail robot. The displacement parameter represents the track deformation.

[0044] This step is executed after obtaining the basic parameters of the handled object, and is used to determine the control parameters of the track operation. Specifically, the system uses the weight, material, and position parameters of the handled object as index conditions to search for matching historical operation records in the preset track database. Through pattern recognition algorithms, the matching results are analyzed to extract the optimal track vibration parameters, operating speed parameters, and track displacement parameters as the control benchmarks for this task.

[0045] In some embodiments, data matching can be achieved in the following ways: Optionally, adopt a parameter matching method based on deep learning, construct a neural network model to establish the mapping relationship between the parameters of the handled object and the track parameters, and achieve fast and accurate matching; Optionally, use a rule-based expert system for parameter matching, and calculate the optimal track operation parameters according to the preset empirical rules. It can be understood that other intelligent algorithms can also be used to achieve parameter matching, which are not limited here.

[0046] S103. Extract the target points in the track path of the ground track, and calculate the handling path data of the ground track robot on the ground track based on the handling parameters of the target points. The target points include the picking point, the transition point, and the placing point.

[0047] Among them, the track path refers to the spatial trajectory curve of the ground track robot's movement. The target point represents the key position node in the track path. The picking point refers to the initial position where the robot grabs the handled object. The transition point is used to represent the intermediate turning position in the path. The placing point refers to the final target position of the handled object. The handling parameters include the spatial coordinates, attitude angles, and motion constraint conditions of the target points. The handling path data refers to the complete data set describing the robot's movement trajectory.

[0048] After obtaining the track operation parameters, it is necessary to plan the specific movement path. Specifically, the system first extracts the spatial coordinates of the picking point, the transition point, and the placing point from the track path. Then, according to the positional relationship and motion constraints of each target point, a smooth and continuous movement path is generated using the trajectory planning algorithm. During the planning process, the robot's kinematic constraints, dynamic constraints, and environmental constraints are considered to ensure the feasibility of the path. Finally, the planning result is converted into a standard path data format, including the position sequence, speed sequence, and acceleration sequence.

[0049] In some embodiments, path planning can be achieved in various ways: Optionally, first establish a cost function for path planning, including evaluation indicators such as path length, smoothness, and safety, then use the A* algorithm to search for the optimal path, and finally use cubic spline interpolation to generate a continuous trajectory; Optionally, construct a virtual force field based on the artificial potential field method, obtain the obstacle avoidance path through iterative calculation, then use quintic polynomial interpolation to generate a trajectory curve with continuous speed, and finally perform dynamic optimization to obtain the final path. It can be understood that other intelligent planning algorithms can also be used to achieve path generation, which is not limited here.

[0050] This step specifically includes:

[0051] Obtain the initial positions of the picking point, the transition point, and the placing point;

[0052] Adjust the initial positions of the picking point, the transition point, and the placing point according to the track vibration parameters to obtain the adjusted positions;

[0053] Calculate the movement trajectory of the ground track robot between each point according to the displacement parameters;

[0054] Calculate the speed change sequence of the ground track robot during the movement process according to the speed parameters;

[0055] Convert the adjusted positions, the movement trajectory, and the speed change sequence into the ground track robot control instruction format to generate the handling path data.

[0056] Among them, the initial position represents the original coordinate value of the target point in the space coordinate system. The orbital vibration parameters refer to the vibration characteristic indexes of the ground track during operation, including amplitude and frequency. The displacement parameter is used to represent the position offset caused by the deformation of the ground track. The speed parameter refers to the speed limit condition during the movement process. The adjusted position represents the coordinate of the target point after vibration compensation. The motion trajectory refers to the motion path curve of the robot in space. The speed change sequence is used to describe the speed change law during the movement process. The control instruction format refers to the standard data format recognizable by the robot control system.

[0057] For the process of target point position adjustment and trajectory planning, the specific implementation method is as follows: First, obtain the initial three-dimensional coordinates of the picking point, transition point, and placing point through the vision positioning system, and the coordinate value accuracy reaches 0.1 mm. Then, perform position compensation based on the orbital vibration parameters, and the compensation amount is calculated through the elastic deformation model. The compensated position can effectively suppress the vibration influence. Next, use the cubic spline interpolation method to generate a continuous trajectory curve between each point, and the trajectory curve satisfies the constraints of position continuity and speed continuity. Based on the trajectory curve, adopt the trapezoidal speed planning algorithm to generate a speed sequence, including an acceleration section, a constant speed section, and a deceleration section, and the speed value does not exceed the preset limit. Finally, pack the adjusted position data, trajectory curve parameters, and speed sequence into robot control instructions, and the instruction format includes position instructions, speed instructions, and timing information. For example, the position instruction uses the format {x, y, z, rx, ry, rz} to represent the spatial position and attitude, the speed instruction uses the format {vx, vy, vz, wx, wy, wz} to represent the linear velocity and angular velocity, and the timing information uses the relative timestamp to represent the execution order.

[0058] S104. Send the handling path data to the ground track robot. After the ground track robot executes the handling path data, measure the real-time force parameters and real-time position parameters of the ground track robot at the picking point, the transition point, and the placing point.

[0059] Among them, the handling path data refers to the complete trajectory information describing the movement of the robot. The real-time force parameters represent the force state of the robot during the movement process, including the force and moment components in three directions. The real-time position parameters refer to the actual spatial position and attitude angle of the robot at the key points. The data transmission adopts the industrial real-time communication protocol to ensure the real-time and reliable transmission of instructions.

[0060] This step is executed after path planning is completed and is used to implement path execution and status monitoring. Specifically, the system first sends the handling path data to the controller of the ground rail robot via the industrial Ethernet. When the robot executes path movement, force sensors and position sensors are installed at the picking point, transition point, and placing point to collect the force state and position state of the robot in real time. After the sensor data is filtered and coordinate-transformed, state parameters in a standard format are formed for subsequent trajectory optimization.

[0061] In some embodiments, data transmission and status monitoring can be implemented in various ways: Optionally, the path data is first packaged into a standard message format and transmitted to the robot controller in real time via the EtherCAT bus. Then, multi-threaded parallel acquisition of force sensor and position sensor data is started, and finally, data synchronization and timestamp alignment are performed. Optionally, the OPCUA protocol is used to establish a data communication link to issue path instructions in real time and feedback the execution status. At the same time, a distributed sensor network is used to collect status parameters, and a unified state estimate is obtained through a data fusion algorithm. It can be understood that other communication methods and status monitoring methods can also be adopted, which are not limited herein.

[0062] S105: Input the real-time force parameter and the real-time position parameter into a preset evaluation model to obtain an orbit correction parameter and a position correction parameter.

[0063] Among them, the preset evaluation model refers to a mathematical model used to evaluate the motion state of the robot and generate a correction strategy. The real-time force parameter refers to the current force state data of the robot. The real-time position parameter is used to represent the spatial position and attitude information of the robot. The orbit correction parameter represents the adjustment amount of the orbit motion parameters. The position correction parameter refers to the compensation amount for the position and attitude of the robot. The input of the evaluation model includes state parameters, the output includes correction parameters, and the intermediate process includes state evaluation and parameter optimization.

[0064] This step is executed after obtaining the real-time state of the robot and is used to generate a motion correction strategy. Specifically, the system preprocesses the collected force parameters and position parameters, including data standardization, outlier detection, and feature extraction. Then, the processed data is input into a preset evaluation model, and the model analyzes the stability and accuracy of the motion state based on machine learning algorithms and calculates the optimal correction parameters. The evaluation model considers multiple evaluation indicators, including position accuracy, attitude stability, dynamic response characteristics, etc., and obtains a balanced correction strategy through multi-objective optimization.

[0065] In some embodiments, state evaluation and parameter optimization can be achieved in various ways: Optionally, first construct an evaluation model based on a deep neural network, perform offline training using historical data, establish a mapping relationship from state parameters to correction parameters, then input the current state in real time for online inference, and finally optimize the correction parameters through the backpropagation algorithm; Optionally, use the fuzzy logic control method to construct an evaluation system, design a fuzzy rule base to describe expert experience, obtain a preliminary correction amount through fuzzy inference, and then use the genetic algorithm to globally optimize the correction parameters. It can be understood that other intelligent algorithms can also be used to achieve state evaluation and parameter optimization, which are not limited herein.

[0066] Before step 105, the following steps are further included:

[0067] Collect historical force parameters of the ground track during multiple operations and historical position parameters of the ground track robot.

[0068] Extract the historical force parameters to obtain force characteristics, and extract the historical position parameters to obtain position characteristics.

[0069] Train a neural network model based on the force characteristics and the position characteristics to obtain the preset evaluation model. The neural network model includes a first sub-network for generating track correction parameters and a second sub-network for generating position correction parameters.

[0070] Among them, the historical force parameters represent the mechanical data recorded by the ground track system during previous operations, including force components (Fx, Fy, Fz) in three directions and torque components (Mx, My, Mz) in three directions. The historical position parameters refer to the six-degree-of-freedom pose data during the actual movement of the robot, including position coordinates (x, y, z) and Euler angles (α, β, γ). The force characteristics are used to represent the statistical and dynamic characteristics of the force data. The position characteristics represent the spatial distribution and temporal characteristics of the position data. The neural network model is a deep learning model that maps input features to correction parameters. The first sub-network maps the force characteristics to track correction amounts (δx, δy, δz). The second sub-network maps the position characteristics to position correction amounts (Δx, Δy, Δz, Δα, Δβ, Δγ).

[0071] The establishment process of the preset evaluation model mainly includes three stages: data collection, feature extraction and model training. First, the ATIF / TSensor six-dimensional force / torque sensor array is installed at the key nodes of the ground rail system, with a sampling frequency of 1kHz, a force measurement range of ±1000N, a torque measurement range of ±100Nm, and a resolution of 0.1N / 0.01Nm; at the same time, the LeicaAT960 laser tracker is used to record the robot position, with a sampling frequency of 100Hz and a measurement accuracy of ±0.02mm / m. In the load range of 10kg to 100kg, different working conditions are set with a step size of 10kg. Each working condition is tested for 100 cycles, each test lasts for 10 minutes, and a total of 120 hours of operation data are obtained. The raw data is low-pass filtered at 50Hz to remove high-frequency noise, the 3σ criterion is used to eliminate outliers, and linear interpolation is used to fill in missing values. Then, the historical force data is feature extracted, and statistical features such as mean μF, standard deviation σF, skewness SF, and kurtosis KF are calculated. The main frequency f0 and its amplitude A0, subharmonic frequency f1, f2 and amplitude A1, A2 are obtained through FFT transformation. The short-time Fourier transform is calculated to obtain the time-frequency feature matrix T(f, t); the spatial distribution features such as position mean μP, variance matrix ΣP, main direction vector v1, v2, v3, and motion features such as velocity sequence V(t), acceleration sequence a(t), and angular velocity sequence ω(t) are calculated for the historical position data. Finally, a two-branch neural network is constructed. The first branch contains 4 convolutional layers and 2 fully connected layers to process 48-dimensional force features, and the second branch contains 1 LSTM layer and 2 fully connected layers to process 36-dimensional position features. The two branches are fused through the attention mechanism and connected to 3 fully connected layers to generate track correction and position correction parameters respectively. The network was trained using a composite loss function that included mean square error loss, parameter regularization term, and smoothness constraint term. The stochastic gradient descent method was used for optimization. The learning rate adopted a cosine annealing strategy and was trained for 200 rounds on 80% of the data. The final model achieved performance indicators of less than 0.1mm average error in orbit correction and less than 0.05mm / 0.01° average error in position correction on the test set.

[0072] S106: Send a correction instruction including the track correction parameter and the position correction parameter to the ground rail robot.

[0073] The correction instruction refers to a control command containing track correction parameters and position correction parameters. Track correction parameters refer to the adjustment amount of track motion characteristics. Position correction parameters are used to indicate the compensation amount for the robot position and posture. The instruction is sent using a real-time communication protocol to ensure the timeliness of the correction. The ground track robot refers to the executor of the correction instruction.

[0074] This step is executed after generating the correction parameters and is used to achieve motion correction. Specifically, the system first encapsulates the orbit correction parameters and the position correction parameters into control instructions in a standard format, including the correction amount, execution timing, and safety constraints. Then, the correction instructions are sent to the controller of the ground rail robot through an industrial real-time network. After receiving the instructions, the controller parses and validates them to ensure that the correction parameters are within the safe range, and then executes the correction actions according to the predetermined strategy to achieve real-time adjustment of the robot's motion.

[0075] In some embodiments, the generation and transmission of the correction instructions can be achieved in various ways: Optionally, first convert the correction parameters into control quantities in the robot joint space, generate motion instructions in a standard format, then send them to the robot controller through real-time Ethernet, and finally monitor the instruction execution status and record the correction effect; Optionally, adopt a hierarchical control architecture, generate a correction strategy at the upper layer, convert it into correction instructions executable by the lower-layer controller, ensure the real-time nature of the instructions through a multi-buffer mechanism, and at the same time establish an emergency handling mechanism to handle communication exceptions. It can be understood that other control methods can also be adopted to achieve motion correction, which is not limited here.

[0076] The following further describes the more specific process of the method provided in this embodiment. Please refer to Figure 2 , which is another schematic flowchart of the multi-degree-of-freedom control method for the ground rail robot in the embodiment of the present application.

[0077] S201. Send the correction instructions including the orbit correction parameters and the position correction parameters to the ground rail robot.

[0078] It can be understood that this step is similar to step S106 and will not be elaborated here.

[0079] S202. Obtain the real-time image data of the ground rail robot, and identify the pose parameters of the carried object from the real-time image data.

[0080] The real-time image data refers to the image information of the carried object collected in real time by an industrial camera installed on the ground rail robot, including RGB color images or depth images. The pose parameters are a set of parameters describing the position and orientation of the carried object in three-dimensional space, including translation parameters (x, y, z coordinates) and rotation parameters (roll, pitch, yaw angles).

[0081] This step is specifically implemented as follows: First, an industrial camera continuously captures images of the transported object at a preset frequency (such as 30 frames per second). Then, computer vision algorithms are used to preprocess the images, including image denoising, distortion correction, etc. Next, a target detection algorithm (such as Faster R-CNN) is used to locate the position of the transported object in the image and extract its contour features. Finally, based on the spatial distribution of the contour feature points, the PnP algorithm is used to calculate the six-degree-of-freedom pose parameters of the transported object. The entire processing process is carried out in real time to ensure the acquisition of the latest pose information.

[0082] S203. When the pose parameter exceeds the preset pose threshold, calculate the pose adjustment amount according to the track correction parameter and the position correction parameter.

[0083] The preset pose threshold is determined according to the accuracy requirements of the handling task. The position deviation threshold is ±5 mm in each of the x, y, and z directions, and the angle deviation threshold is ±2° for roll, pitch, and yaw respectively. The track correction parameters include the deformation amount, vibration amplitude, and stiffness coefficient of each section of the track, which are obtained by real-time measurement using sensors on the track. The position correction parameters include the position compensation amount, pose compensation amount, and dynamic response characteristics, which are calculated by the control system based on historical operation data.

[0084] When calculating the pose adjustment amount, first establish a deformation model based on the track state, considering the effects of elastic deformation and temperature deformation of the track. At the same time, establish a position constraint model, including kinematic constraints of the robotic arm and workspace constraints. Combine these two models and use the least squares method to calculate the optimal adjustment amount that brings the pose parameter back within the threshold range. The calculation results include displacement adjustment amounts (Δx, Δy, Δz) in three directions and angle adjustment amounts (Δroll, Δpitch, Δyaw) in three directions. Finally, verify whether the adjustment amount meets the motion range limit of the robotic arm.

[0085] S204. Determine the execution speed of the pose adjustment based on the weight parameter and the material parameter.

[0086] In addition to the mass value of the transported object, the weight parameter also includes the center of gravity position and the moment of inertia. The material parameter describes the physical properties of the transported object, including the material type, hardness value, surface friction coefficient, elastic modulus, and impact resistance. The execution speed is divided into translational speed and rotational speed, which need to be reasonably set according to the characteristics of the transported object to ensure the safety of the adjustment process.

[0087] The speed determination process adopts a fuzzy control method: normalizing the weight parameter to the range of 0 - 1, and comprehensively quantifying the material properties as a safety factor. According to the empirical rule, the greater the weight and the more fragile the material, the lower the execution speed; the lighter the weight and the stronger the material, the higher the execution speed. The linear speed calculation formula is v = vmax×(1 - w)×h, and the angular speed calculation formula is ω = ωmax×(1 - w)×h, where vmax and ωmax are the maximum allowable speeds, w is the normalized weight value, and h is the material safety factor. The calculated speed also needs to satisfy the acceleration constraint, generating a smooth speed trajectory through speed planning, and finally converting it into the robot joint speed command.

[0088] S205. Send the pose adjustment instruction containing the pose adjustment amount and the execution speed to the gantry robot.

[0089] The pose adjustment instruction is a data packet containing displacement adjustment amounts (Δx, Δy, Δz), angle adjustment amounts (Δroll, Δpitch, Δyaw), and corresponding execution speed parameters (vx, vy, vz, ωroll, ωpitch, ωyaw). The gantry robot includes a motion control unit, a communication unit, and an actuator. The motion control unit is responsible for parsing the instruction and generating a motion trajectory. The communication unit uses industrial Ethernet for data transmission. The actuator includes a servo motor and a driver.

[0090] This step sends the pose adjustment instruction to the motion control unit of the gantry robot through industrial Ethernet. The motion control unit first converts the adjustment amounts and speed parameters into motion parameters in the joint space, including the target positions and speeds of each joint. Then, based on the quintic polynomial, trajectory planning is performed to generate the position, speed, and acceleration curves of each joint, ensuring that the motion trajectory is smooth and continuous. The trajectory planning result is converted into a servo control instruction and sent to the servo driver of each joint in real time through the EtherCAT bus. After receiving the instruction, the servo driver controls the servo motor to perform the pose adjustment action according to the planned trajectory. The entire communication process uses a real-time Ethernet protocol, and the communication cycle is 1 ms, ensuring the real-time performance and synchronization of the instruction.

[0091] S206. When the pose adjustment instruction starts to be executed, collect the real-time force parameters of the gantry.

[0092] The real-time force parameters refer to the forces and torques borne by the gantry during the robot's movement, including the force components (Fx, Fy, Fz) and torque components (Mx, My, Mz) in the x, y, and z directions. A six-axis force / torque sensor array is installed on the gantry. The sampling frequency of each sensor is 1 kHz, the measurement range is force ±1000 N, torque ±100 Nm, and the measurement accuracy is 0.1% of the full scale.

[0093] This step starts the data acquisition of the force / torque sensor simultaneously when the attitude adjustment instruction begins to execute. After the sensor converts the analog signal into a digital signal, it is transmitted to the data acquisition unit via the CAN bus. The data acquisition unit filters and performs coordinate transformation on the original data to obtain the six-dimensional force / torque data in the ground rail base coordinate system. To improve data reliability, the data of multiple sensors are weighted and fused, and the weight coefficients are determined according to the distance from the sensor to the action point. The processed force parameters are updated in real time at a frequency of 100 Hz for subsequent safety monitoring.

[0094] S207. When the real-time force parameter is greater than the preset safety threshold, a pause instruction is sent to the ground rail robot to make the ground rail robot pause the attitude adjustment.

[0095] The preset safety threshold is the force and torque limit values determined according to the load-bearing capacity and structural strength of the ground rail, including the safety thresholds of force (Fx_max, Fy_max, Fz_max) and the safety thresholds of torque (Mx_max, My_max, Mz_max). The pause instruction is an emergency control instruction that includes a deceleration parameter and a stop mode parameter. The pause process of the ground rail robot needs to ensure smooth deceleration to avoid impact caused by sudden stop.

[0096] This step first compares the six components of the real-time force parameter with the corresponding safety thresholds. When any component exceeds the safety threshold, the control system immediately generates a pause instruction, which includes the maximum allowable deceleration (2 m / s²) and a soft stop mode flag. The pause instruction is sent to the motion control unit of the ground rail robot via the industrial Ethernet. After receiving the instruction, the control unit immediately stops the current trajectory planning and regenerates a deceleration trajectory. The deceleration trajectory uses a quadratic curve to ensure that each joint decelerates smoothly until it stops. During the deceleration process, the servo driver maintains the torque output to prevent the robot from falling out of control. When the robot stops completely, the servo system enters the position holding mode and waits for subsequent processing instructions.

[0097] S208. When the real-time force parameter returns to within the preset safety threshold range, the execution speed is reduced to the preset execution speed threshold.

[0098] The preset safety threshold range refers to the safe intervals of force and torque limits, including the safe intervals of force ([0.7×Fx_max, 0.8×Fx_max], [0.7×Fy_max, 0.8×Fy_max], [0.7×Fz_max, 0.8×Fz_max]) and the safe intervals of torque ([0.7×Mx_max, 0.8×Mx_max], [0.7×My_max, 0.8×My_max], [0.7×Mz_max, 0.8×Mz_max]). The preset execution speed threshold is the deceleration ratio value of the original execution speed. The linear velocity threshold is 0.6 times the original speed, and the angular velocity threshold is 0.5 times the original speed.

[0099] The control system continuously monitors the six-dimensional force / torque data. When it detects that all components have returned to the safe interval, it recalculates the execution speed. The new linear velocity value v_new = 0.6×v_original, and the new angular velocity value ω_new = 0.5×ω_original. The control system generates a motion instruction containing the new speed parameters and sends it to the ground rail robot through the industrial Ethernet. After receiving the instruction, the motion control unit uses a trapezoidal speed curve to generate an acceleration trajectory and gradually accelerates the robot from the stationary state to the new execution speed. A fixed acceleration (1m / s²) is used during the acceleration process to ensure a smooth transition. After the robot reaches the new speed, it continues to perform the remaining attitude adjustment actions.

[0100] S209. Determine the track vibration threshold based on the weight parameter and the material parameter.

[0101] The track vibration threshold includes a displacement amplitude threshold (unit: mm) and an acceleration amplitude threshold (unit: m / s²). The weight parameter includes the mass value of the transported object (unit: kg) and the mass distribution parameter. The material parameter includes the elastic modulus of the material (unit: GPa), the damping ratio, and the seismic performance index. These parameters jointly determine the dynamic response characteristics of the ground rail system during the transportation process.

[0102] This step first calculates the reference vibration threshold according to the weight parameter. The displacement amplitude reference value A_base = 10 / m, and the acceleration amplitude reference value a_base = 20 / m, where m is the mass of the transported object (kg). Then, it calculates the correction coefficient according to the material parameter. The correction coefficient k = E×ζ / E_ref, where E is the elastic modulus of the material, ζ is the damping ratio, and E_ref is the reference elastic modulus (taking 200GPa). The final displacement amplitude threshold A_limit = A_base×k, and the acceleration amplitude threshold a_limit = a_base×k. The calculation of the vibration threshold takes into account the mass effect of the transported object and the material characteristics and is used for subsequent vibration monitoring.

[0103] S210. During the operation of the ground rail robot, collect the real-time rail vibration parameters of each rail section.

[0104] The real-time rail vibration parameters include displacement amplitude (unit: mm), acceleration amplitude (unit: m / s²), main vibration frequency (unit: Hz), and vibration waveform characteristics. A rail section refers to a physical segment in the ground rail system, with each segment being 2 meters long. An acceleration sensor and a displacement sensor are installed on each rail section. The sampling frequency of the acceleration sensor is 2 kHz, and the measuring range is ±50 m / s²; the sampling frequency of the displacement sensor is 1 kHz, and the measuring range is ±10 mm.

[0105] The sensor array monitors the rail vibration in real time. The collected raw data is amplified and filtered by the signal conditioning circuit. The data acquisition unit performs digital conversion on the processed signal and calculates the vibration parameters. The displacement amplitude is directly measured by the displacement sensor, the acceleration amplitude is measured by the acceleration sensor, and the main vibration frequency is obtained by performing FFT analysis on the acceleration signal. The vibration waveform characteristics are obtained by extracting the time-domain and frequency-domain characteristics of the signal. The processed vibration parameters are updated at a frequency of 50 Hz to form the vibration status data stream of each rail section. These data are used for evaluating the dynamic performance of the rail and early warning analysis.

[0106] S211. Compare the real-time rail vibration parameters with the maximum allowable rail vibration threshold, mark the rail sections that exceed the maximum allowable rail vibration threshold, and obtain the marked rail sections.

[0107] The real-time rail vibration parameters include displacement amplitude (A), acceleration amplitude (a), and main vibration frequency (f). The maximum allowable rail vibration threshold includes displacement amplitude threshold (A_limit), acceleration amplitude threshold (a_limit), and frequency range threshold (f_range). The marked rail sections record the vibration overrun status through a binary flag bit (status_flag), including the rail section number (section_id), the type of overrun parameter (exceed_type), and the degree of overrun (exceed_level).

[0108] The control system evaluates the vibration parameters of each track section at a frequency of 50 Hz. In the evaluation process, the displacement amplitude ratio RA = A / A_limit, the acceleration amplitude ratio Ra = a / a_limit, and the frequency deviation ratio Rf = |f - f_center| / f_range are first calculated, where f_center is the expected vibration frequency. When any of the ratios is greater than 1, the status_flag of the corresponding track section is set to 1, and at the same time, the exceed_type (1 indicates displacement overrun, 2 indicates acceleration overrun, 3 indicates frequency overrun) and exceed_level (equal to the maximum overrun ratio) are recorded. The marked information forms a structured data packet, which contains four fields: section_id, status_flag, exceed_type, and exceed_level, and is used for subsequent early warning processing.

[0109] S212. Send the position of the marked track section to the target client.

[0110] The position information of the marked track section includes the start coordinate (start_pos), end coordinate (end_pos), and center coordinate (center_pos) of the track section. The target client includes on-site control terminals and remote monitoring terminals, communicates using the TCP / IP protocol, and data transmission uses the JSON format. The client has functions of data reception, parsing, and display.

[0111] The control system packages the position information and vibration parameters of the marked track section into a JSON data packet. The data packet structure includes: basic track section information (section_info), position information (position_info), vibration parameters (vibration_info), and overrun information (exceed_info). The data packet is sent to the target client through the industrial Ethernet, and the transmission protocol uses TCP / IP to ensure the reliability of data transmission. After receiving the data packet, the client performs JSON parsing, extracts the track section position information and vibration parameters, and displays the position and status of the overrun track section graphically on the human-machine interface. The display content includes the position of the track section in three-dimensional space, the real-time values of the vibration parameters, and the color identification of the overrun degree.

[0112] S213. Process the real-time image data using AR-assisted visual recognition technology to obtain environmental characteristic parameters.

[0113] The AR-assisted visual recognition technology is a technical system that integrates augmented reality and computer vision, including a camera calibration module (for establishing the internal and external parameter models of the camera), a feature extraction module (for extracting key features of the image), and a spatial registration module (for realizing virtual-real registration). The real-time image data includes RGB images (with a resolution of 4096×3072 pixels and a frame rate of 30fps) and depth images (with a resolution of 1280×720 pixels and a depth accuracy of 0.1mm). The environmental feature parameters include three categories: spatial geometric features (plane equation system, edge equation system, three-dimensional coordinates of corner points), lighting conditions (average brightness value, contrast value, main light source direction vector), and occlusion conditions (set of boundary points of the occlusion area, size parameters of the occluding object).

[0114] The image processing process adopts a multi-stage serial processing method: In the first stage, the RGB image is preprocessed, Gaussian filtering is used to remove noise, histogram equalization is adopted to enhance the image contrast, and distortion correction is performed based on the camera calibration parameters. In the second stage, SIFT feature points are extracted from the processed image, and each feature point contains position coordinates, scale, direction, and a 128-dimensional feature descriptor. In the third stage, the feature points are corresponded to the depth image to construct the three-dimensional coordinates of the feature points, forming feature point cloud data. In the fourth stage, the ICP algorithm is used for point cloud registration, and the position matrix T and rotation matrix R of the camera in the world coordinate system are iteratively calculated. In the fifth stage, environmental features are extracted based on the registration result: the plane equation system Ax + By + Cz + D = 0 is fitted through the RANSAC algorithm; the Canny operator is used to detect edges and fit the straight line equation system; the coordinates of feature corner points are obtained through Harris corner detection. In the sixth stage, the image brightness distribution is analyzed, the average brightness value L_avg and contrast value C are calculated, and the main light source direction vector is determined through brightness gradient analysis. In the seventh stage, the depth image's discontinuity is used to detect the occlusion area, the set of occlusion boundary points {(x_i, y_i, z_i)} is extracted, and the bounding box parameters of the occluding object are calculated.

[0115] S214. Adjust the orbit correction parameter and the position correction parameter according to the environmental feature parameter.

[0116] The orbit correction parameter includes the orbit geometric correction amount (δx, δy, δz) and the attitude correction amount (δroll, δpitch, δyaw). The position correction parameter includes the spatial position compensation amount (Δx, Δy, Δz) and the attitude compensation amount (Δroll, Δpitch, Δyaw). The adjustment of these parameters is dynamically updated based on the environmental feature parameters.

[0117] The adjustment process is divided into two parts: orbit correction and position correction. For orbit correction, the detected planar features are first used to establish a ground reference plane, and the deviation between the actual orbit plane and the reference plane is calculated to obtain the orbit geometric correction amounts: δx = k1×(x_actual - x_ref), δy = k1×(y_actual - y_ref), δz = k1×(z_actual - z_ref), where k1 is the correction coefficient (taking the value of 0.8), (x_actual, y_actual, z_actual) are the actual plane parameters, and (x_ref, y_ref, z_ref) are the reference plane parameters. The attitude correction amounts are calculated through edge features: δroll = k2×(θx_actual - θx_ref), δpitch = k2×(θy_actual - θy_ref), δyaw = k2×(θz_actual - θz_ref), where k2 is the attitude correction coefficient (taking the value of 0.6), and θ represents the inclination angles in each direction. For position correction, the compensation amounts are adjusted according to the occlusion situation: when occlusion is detected, the position compensation amounts increase the safety margin according to the size of the occluding object, Δx = Δx_base + d_safe, Δy = Δy_base + d_safe, Δz = Δz_base + d_safe, where d_safe is the safety distance (taking the value of 100 mm). At the same time, the illumination conditions affect the attitude compensation: in areas with uneven illumination (contrast C > 0.7), the attitude compensation amounts are increased to improve stability, Δroll = k3×Δroll_base, Δpitch = k3×Δpitch_base, Δyaw = k3×Δyaw_base, where k3 is the illumination compensation coefficient (taking the value of 1.2).

[0118] The multi-degree-of-freedom control system of the ground-rail robot in the embodiment of the present invention application will be described from the perspective of hardware processing. Please refer to Figure 3 , which is a schematic structural diagram of an entity device of the multi-degree-of-freedom control system of the ground-rail robot in the embodiment of the present application.

[0119] It should be noted that Figure 3 the structure of the multi-degree-of-freedom control system of the ground-rail robot shown is only an example, and should not bring any limitations to the functions and application scope of the embodiments of the present invention.

[0120] As shown in Figure 3As shown, the multi-degree-of-freedom control system of the floor track robot includes a Central Processing Unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the Read-Only Memory (ROM) 302 or the program loaded from the storage section 308 into the Random Access Memory (RAM) 303, such as executing the method described in the above embodiments. In the RAM 303, various programs and data required for system operation are also stored. The CPU 301, ROM 302, and RAM 303 are connected to each other via a bus 304. An Input / Output (I / O) interface 305 is also connected to the bus 304.

[0121] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a Liquid Crystal Display (LCD), an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A driver 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the driver 310 as needed so that a computer program read from it can be installed into the storage section 308 as needed.

[0122] Specifically, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication section 309, and / or installed from the removable medium 311. When the computer program is executed by the Central Processing Unit (CPU) 301, various functions defined in the present invention are executed.

[0123] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0124] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. Among them, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the above module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings.

[0125] Specifically, the multi-degree-of-freedom control system of the ground rail robot in this embodiment includes a processor and a memory. A computer program is stored on the memory, and when the computer program is executed by the processor, it implements the multi-degree-of-freedom control method of the ground rail robot provided in the above embodiment.

[0126] On the other hand, the present invention also provides a computer-readable storage medium, which may be included in the multi-degree-of-freedom control system of the ground rail robot described in the above embodiment; or it may exist alone and not be assembled into the multi-degree-of-freedom control system of the ground rail robot. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of a multi-degree-of-freedom control system of the ground rail robot, the multi-degree-of-freedom control system of the ground rail robot is enabled to implement the multi-degree-of-freedom control method of the ground rail robot provided in the above embodiment.

[0127] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

[0128] As used in the foregoing embodiments, depending on the context, the term "when" may be construed to mean "if" or "after" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "upon determining" or "if (the stated condition or event) is detected" may be construed to mean "if determined" or "in response to determining" or "when (the stated condition or event) is detected" or "in response to detecting (the stated condition or event)".

[0129] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the foregoing embodiments can be implemented by a computer program instructing relevant hardware. The program can be stored in a computer-readable storage medium. When the program is executed, it may include the processes of the foregoing method embodiments. The foregoing storage media include: various media such as ROM or random access memory RAM, magnetic disks, or optical discs that can store program codes.

Claims

1. A multi-degree-of-freedom control method for a ground rail robot, characterized in that: Applied to a multi-degree-of-freedom control system of a ground rail robot, the multi-degree-of-freedom control system of the ground rail robot comprises a ground rail robot and a ground rail, and the method comprises: Acquire the task data of the transported object, extract the characteristic quantities in the task data of the transported object to obtain the weight parameter, material parameter and position parameter of the transported object; Matching the track operation data of the ground rail in a preset track database according to the weight parameter, the material parameter and the position parameter, wherein the track operation data includes a track vibration parameter, a speed parameter and a displacement parameter; Get the initial positions of the pick-up point, transition point, and placement point; adjusting the initial positions of the pick-up point, the transition point and the placement point according to the track vibration parameter to obtain an adjusted position; Calculate the motion trajectory of the ground rail robot between various points according to the displacement parameters; Calculating a speed change sequence of the ground rail robot during movement according to the speed parameter; Converting the adjustment position, the motion trajectory and the speed change sequence into a control instruction format of a ground rail robot to generate transport path data, wherein the target points include a pick-up point, a transition point and a placement point; The transport path data is sent to the ground rail robot, and after the ground rail robot executes the transport path data, the real-time force parameters and real-time position parameters of the ground rail robot are measured at the pick-up point, the transition point and the placement point; Collect historical force parameters of the ground rail during multiple operations and historical position parameters of the ground rail robot; Extracting the historical force parameters to obtain force characteristics, and extracting the historical position parameters to obtain position characteristics; A neural network model is trained based on the force characteristics and the position characteristics to obtain a preset evaluation model, wherein the neural network model includes a first subnetwork for generating track correction parameters and a second subnetwork for generating position correction parameters, the historical force parameters include force components (Fx, Fy, Fz) in three directions and moment components (Mx, My, Mz) in three directions, and the historical position parameters include position coordinates (x, y, z) and Euler angles (α, β, γ); Inputting the real-time force parameter and the real-time position parameter into a preset evaluation model to obtain a track correction parameter and a position correction parameter; A correction instruction including the track correction parameter and the position correction parameter is sent to the ground rail robot.

2. The method according to claim 1, characterized in that After the step of sending the correction instruction including the track correction parameter and the position correction parameter to the ground rail robot, the method further includes: Acquiring real-time image data of the ground rail robot, and identifying posture parameters of the transported object from the real-time image data; When the attitude parameter exceeds a preset attitude threshold, calculating an attitude adjustment amount according to the trajectory correction parameter and the position correction parameter; determining an execution speed of posture adjustment based on the weight parameter and the material parameter; A posture adjustment instruction is sent to the ground rail robot, wherein the posture adjustment instruction includes the posture adjustment amount and the execution speed.

3. The method according to claim 2, characterized in that After the step of sending the posture adjustment instruction including the posture adjustment amount and the execution speed to the ground rail robot, the method further includes: When the attitude adjustment instruction starts to be executed, real-time force parameters of the ground rail are collected; When the real-time force parameter is greater than a preset safety threshold, a pause instruction is sent to the ground rail robot so that the ground rail robot pauses posture adjustment; When the real-time force parameter is restored to within the preset safety threshold range, the execution speed is reduced to a preset execution speed threshold.

4. The method according to claim 1, characterized in that: After the step of inputting the real-time force parameter and the real-time position parameter into a preset evaluation model to obtain a track correction parameter and a position correction parameter, the method further includes: determining a rail vibration threshold based on the weight parameter and the material parameter; During the operation of the ground rail robot, real-time track vibration parameters of each track segment are collected; Comparing the real-time track vibration parameter with a maximum allowable track vibration threshold, marking a track segment that exceeds the maximum allowable track vibration threshold, and obtaining a marked track segment; The position of the marked track segment is sent to the target client.

5. The method according to claim 2, characterized in that: After the step of inputting the real-time force parameter and the real-time position parameter into a preset evaluation model to obtain a track correction parameter and a position correction parameter, the method further includes: The real-time image data is processed using AR-assisted visual recognition technology to obtain environmental characteristic parameters; The trajectory correction parameter and the position correction parameter are adjusted according to the environmental characteristic parameter.

6. A multi-degree-of-freedom control system for a ground-rail robot, characterized in that: The multi-degree-of-freedom control system of the ground rail robot includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program codes, the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the multi-degree-of-freedom control system of the ground rail robot to execute the method described in any one of claims 1-5.

7. A computer-readable storage medium comprising instructions, characterized in that: When the instruction runs on a rail robot multi-degree-of-freedom control system, the rail robot multi-degree-of-freedom control system executes the method according to any one of claims 1 to 5.

8. A computer program product, characterized in that When the computer program product runs on a rail robot multi-degree-of-freedom control system, the rail robot multi-degree-of-freedom control system is enabled to execute the method according to any one of claims 1 to 5.

Images