AGV self-adaptive material docking control method and system and storage medium

Abstract

The invention discloses an AGV adaptive material docking control method and system and a storage medium, and relates to the technical field of automatic logistics. The method comprises the following steps: installing a sensor group on a target AGV; a material sensing data flow is obtained through real-time monitoring of a sensor group, and sensing parameters are obtained through analysis and processing; obtaining an AGV dynamic analysis model according to an adaptive algorithm, and performing docking analysis on the sensing parameters to obtain applicable docking parameters; multi-dimensional characteristic information of the target material is obtained, an optimization control strategy is adopted to carry out optimization analysis on the applicable docking parameters, and target docking parameters are determined; and based on the target docking parameter, controlling the target AGV to carry out adaptive docking control so as to solve the technical problems of material pushing and material collision caused by environment change, inconsistent material height or errors in the material docking process of the AGV and insufficient platform docking precision.

Description

Technical Field

[0001] The present invention relates to the field of automated logistics technology, and in particular to an AGV adaptive material docking control method, system and storage medium. Background Art

[0002] With the continuous development of automated logistics and intelligent manufacturing technologies, automatic guided vehicles (AGVs) are increasingly used in industrial production and logistics systems. AGVs are mainly used to automatically handle materials and realize automatic transportation and docking of materials. However, in actual applications, AGVs often push or bump into materials when docking. This is mainly manifested in errors in the AGV platform itself, or environmental factors causing the AGV and material positions to shift, the docking height to be incorrect, or the platform extension distance to be incorrect during docking. This not only affects production efficiency, but also causes damage to equipment or goods and personnel safety accidents. Summary of the invention

[0003] The present application provides an AGV adaptive material docking control method, system and storage medium, which solve the technical problems of pushing and bumping materials caused by environmental changes, inconsistent material heights or errors, and insufficient platform docking accuracy during the AGV material docking process. By combining adaptive algorithms and optimized control strategies, the docking parameters are dynamically adjusted according to the multi-dimensional characteristic information of the materials, achieving the technical effect of improving the AGV material docking accuracy, reducing the pushing and bumping materials, and improving production efficiency.

[0004] In view of the above problems, the present application provides an AGV adaptive material docking control method, system and storage medium.

[0005] The first aspect of the present application provides an AGV adaptive material docking control method, the method comprising: installing a sensor group on a target AGV, the sensor group comprising a laser ranging sensor, an ultrasonic sensor, an infrared sensor and a visual sensor; acquiring a material perception data stream through real-time monitoring by the sensor group, analyzing and processing the material perception data stream, and obtaining perception parameters of the target AGV and the target material, the perception parameters comprising relative position, height difference and depth; acquiring an AGV dynamic analysis model according to an adaptive algorithm, performing a docking analysis on the perception parameters based on the AGV dynamic analysis model, and obtaining applicable docking parameters; acquiring multi-dimensional characteristic information of the target material, optimizing and analyzing the applicable docking parameters based on the multi-dimensional characteristic information using an optimization control strategy, and determining target docking parameters; and controlling the target AGV to perform adaptive docking control on the target material based on the target docking parameters.

[0006] The second aspect of the present application provides an AGV adaptive material docking control system, the system comprising: a sensor group installation module: installing a sensor group on a target AGV, the sensor group comprising a laser ranging sensor, an ultrasonic sensor, an infrared sensor and a visual sensor; an analysis and processing module: acquiring material perception data streams through real-time monitoring of the sensor group, analyzing and processing the material perception data streams, and obtaining perception parameters of the target AGV and the target material, the perception parameters comprising relative position, height difference and depth; a docking analysis module: acquiring an AGV dynamic analysis model according to an adaptive algorithm, performing docking analysis on the perception parameters based on the AGV dynamic analysis model, and obtaining applicable docking parameters; an optimization analysis module: acquiring multi-dimensional characteristic information of the target material, optimizing and analyzing the applicable docking parameters based on the multi-dimensional characteristic information using an optimization control strategy, and determining target docking parameters; a docking control module: controlling the target AGV to perform adaptive docking control on the target material based on the target docking parameters.

[0007] The third aspect of the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the AGV adaptive material docking control method provided in the present application.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] First, a sensor group is installed on the target AGV, the sensor group includes a laser rangefinder sensor, an ultrasonic sensor, an infrared sensor and a visual sensor, and the sensor group is used to monitor and obtain the material perception data stream in real time, analyze and process the material perception data stream, and obtain the perception parameters of the target AGV and the target material, the perception parameters include relative position, height difference and depth; then, the AGV dynamic analysis model is obtained according to the adaptive algorithm, and the perception parameters are docked and analyzed based on the AGV dynamic analysis model to obtain applicable docking parameters; further, the multi-dimensional characteristic information of the target material is obtained, and the optimization control strategy is used to optimize and analyze the applicable docking parameters based on the multi-dimensional characteristic information to determine the target docking parameters; finally, the target AGV is controlled to perform adaptive docking control on the target material based on the target docking parameters. The technical problems of pushing and bumping materials caused by environmental changes, inconsistent material height or errors, and insufficient platform docking accuracy during the material docking process of AGV are solved. By combining the adaptive algorithm and the optimization control strategy, the docking parameters are dynamically adjusted according to the multi-dimensional characteristic information of the material, so as to achieve the technical effect of improving the docking accuracy of AGV materials, reducing the pushing and bumping materials, and improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A schematic diagram of the process flow of the AGV adaptive material docking control method provided in an embodiment of the present application;

[0012] Figure 2 This is a schematic diagram of the structure of the AGV adaptive material docking control system provided in an embodiment of the present application.

[0013] Explanation of the reference numerals: sensor group installation module 11 , analysis and processing module 12 , docking analysis module 13 , optimization analysis module 14 , docking control module 15 . DETAILED DESCRIPTION

[0014] The present application solves the technical problems of pushing and bumping materials caused by environmental changes, inconsistent material heights or errors, and insufficient platform docking accuracy during the material docking process of AGV by providing an AGV adaptive material docking control method, system and storage medium. By combining adaptive algorithms and optimized control strategies, the docking parameters are dynamically adjusted according to the multi-dimensional characteristic information of the materials, thereby achieving the technical effect of improving the AGV material docking accuracy, reducing the pushing and bumping materials, and improving production efficiency.

[0015] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0016] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0017] Embodiment 1, as Figure 1 As shown, the present application provides an AGV adaptive material docking control method, wherein the method includes:

[0018] A sensor group is installed on the target AGV, wherein the sensor group includes a laser ranging sensor, an ultrasonic sensor, an infrared sensor, and a visual sensor.

[0019] In one embodiment, the system terminal installs a group of sensors on the target AGV, including laser ranging sensors, ultrasonic sensors, infrared sensors and visual sensors; the laser ranging sensor is used to accurately measure the distance to the object to ensure that the AGV can accurately identify the spatial relationship between the material and the object; the ultrasonic sensor can detect in a more complex environment to help the AGV identify the existence of obstacles or materials; the infrared sensor enhances the perception ability of the AGV in a specific environment, especially in low light conditions, by sensing the heat changes of the object; the visual sensor uses image recognition technology to help the AGV locate the material, recognize the posture and make accurate judgments during the docking process through the images captured by the camera; these sensors work together to provide comprehensive perception information to ensure that the AGV can accurately complete the docking task of materials in a complex environment.

[0020] The material perception data stream is acquired through real-time monitoring by the sensor group, and the material perception data stream is analyzed and processed to obtain the perception parameters of the target AGV and the target material, wherein the perception parameters include relative position, height difference and depth.

[0021] In one embodiment, the system terminal collects the perception data stream of the material in real time through the sensor group installed on the AGV. After capturing the spatial information between the material and the AGV, the sensor converts this information into a data stream and transmits it to the system terminal; then, these data streams are analyzed and processed (including denoising, feature extraction, fusion, etc.) to extract key perception parameters, mainly including the relative position between the AGV and the material, the height difference between the material and the AGV platform, and the depth relationship between the material and the platform during the docking process; these perception parameters help the AGV to accurately understand the specific position of the material, the accuracy and depth of the docking, and ensure the accuracy and safety of the material docking operation.

[0022] Furthermore, the obtaining of the perception parameters of the target AGV and the target material includes:

[0023] Perform noise characteristic analysis on the material perception data stream to obtain perception data noise characteristic information; perform denoising on the material perception data stream based on the perception data noise characteristic information to obtain a standard material perception data stream; perform feature extraction and feature fusion on the standard material perception data stream to obtain a material fusion perception feature set; perform key parameter analysis on the material fusion perception feature set to determine the perception parameters of the target AGV and target material.

[0024] Preferably, after obtaining the material perception data stream, the system terminal performs noise characteristic analysis on the collected material perception data stream in order to identify the noise source in the data; in this process, the system terminal uses a noise identification algorithm (such as frequency domain analysis, time domain analysis, etc.) to analyze the data stream and detect noise characteristics that may interfere with material positioning, thereby forming perception data noise characteristic information. These noise characteristic information can be the amplitude and frequency distribution of random noise in the signal, or data fluctuations caused by environmental factors (such as changes in light, temperature, etc.).

[0025] Based on the perception data noise characteristic information extracted in the previous step, the system terminal denoises the material perception data stream, which usually involves the application of denoising algorithms, such as Kalman filtering, mean filtering or wavelet denoising, to remove unnecessary noise and retain valid signals. These algorithms will adjust the filter parameters according to the perception data noise characteristic information (such as noise amplitude, frequency range, etc.) to effectively remove interference signals, thereby obtaining a more accurate standard material perception data stream. This standard material perception data stream can more realistically reflect the spatial relationship between materials and AGVs.

[0026] The denoised standard material perception data stream will be further processed. First, feature extraction will be performed. This step will extract multiple key features from the data stream, such as the boundary, shape, relative position, motion trajectory, etc. of the material. These features may come from different types of sensors, such as the distance information of the laser ranging sensor, the image information of the visual sensor, etc.; then, the system terminal will use feature fusion technology to merge the extracted features, which can be processed through algorithms such as weighted averaging and principal component analysis (PCA) to unify the features of different data sources into a fused perception feature set, so that the system terminal can comprehensively judge the multi-dimensional features of the material, thereby improving the accurate judgment of the material position and status.

[0027] Based on the obtained material fusion perception feature set, key parameter analysis is performed. By analyzing the material fusion perception feature set, the system terminal can extract core parameters related to material docking, such as the relative position between the material and AGV, the height difference between the material and the platform, the docking depth, etc. These parameters will together constitute the perception parameters between the target AGV and the target material. These perception parameters will provide necessary data support for subsequent docking control, ensuring that the AGV can accurately and safely dock with the material.

[0028] Furthermore, the obtaining of the material fusion perception feature set includes:

[0029] According to the sensor group, a multi-sensor feature extraction algorithm set is determined; according to the multi-sensor feature extraction algorithm set, feature extraction is performed on the standard material perception data stream respectively to obtain a multi-sensor perception feature set; and a data fusion algorithm is used to align and fuse the multi-sensor perception feature set to obtain the material fusion perception feature set.

[0030] Optionally, according to the sensor group used (such as laser ranging sensors, ultrasonic sensors, infrared sensors and visual sensors, etc.), the system terminal first selects a set of feature extraction algorithms, which extract relevant information according to the characteristics of each sensor; for example, the laser ranging sensor may use a distance measurement algorithm (such as triangulation, time of flight, etc.) to determine the distance of the object, and the visual sensor may use an image processing algorithm (such as Canny edge detection, Hough transform, etc.) to extract the boundary and shape information of the material; for each sensor, the system terminal will select a suitable algorithm to extract useful features from the perception data.

[0031] After determining the feature extraction algorithm set, the system terminal will apply these algorithms to process the standard material perception data stream respectively, and each sensor will use its own corresponding algorithm to extract features; for example, for a laser ranging sensor, the system terminal will obtain the laser receiving signal sent back by the laser ranging sensor, and extract the time difference between laser emission and reception, and then multiply this time difference by the speed of light. Because the propagation of light signals is bidirectional, the calculated product is divided by 2 to obtain the distance information of the material; the ultrasonic sensor also provides obstacle detection information of the material through the time-of-flight method, and the visual sensor analyzes the image data through Canny edge detection to identify the shape, boundary, size, etc. of the material; through these algorithm processing, the system terminal obtains a set of feature data from each sensor, that is, a multi-sensor perception feature set.

[0032] After obtaining the multi-sensor perception feature set, the system terminal uses a data fusion algorithm to fuse the perception features obtained from different sensors. This process first requires the alignment of the data from each sensor to ensure that the features from different sources can be compared and merged in the same reference frame. The fusion method may include weighted averaging, principal component analysis (PCA) or other data fusion methods. Taking principal component analysis as an example, PCA is used to map multi-dimensional feature data to a low-dimensional space, thereby reducing redundancy and retaining the most important feature information, obtaining a comprehensive material fusion perception feature set, and providing more comprehensive and accurate data support for subsequent docking operations.

[0033] Furthermore, the real-time monitoring and acquisition of material perception data stream by the sensor group includes:

[0034] Acquire application environment information of the target AGV; perform offset calibration on each sensor in the sensor group based on the application environment information to obtain a standard sensor group; set a data perception frequency, use the standard sensor group according to the data perception frequency, and monitor and acquire the material perception data stream in real time.

[0035] Optionally, the system terminal collects relevant information about the working environment of the target AGV and stores this information in the application environment information. The environmental information includes temperature, humidity, light intensity, etc. These factors may affect the performance and measurement accuracy of the sensor and can be obtained through built-in environmental monitoring equipment (such as temperature and humidity sensors, light sensors, etc.).

[0036] Based on the acquired application environment information, the system terminal performs offset calibration on each sensor in the sensor group; for example, temperature changes may cause measurement errors of laser ranging sensors, and changes in light intensity may affect the image recognition accuracy of visual sensors; by inputting the application environment information and the current parameters of each sensor into the measurement error model, the model adjusts the current parameters of each sensor (such as the resolution, reflectivity, response time, etc. of laser ranging sensors, the resolution, focal length, field of view, etc. of visual sensors) according to the learned mapping relationship to adapt it to the current environment, eliminate the deviation caused by environmental factors, and generate calibration parameters for each sensor; then, these calibration parameters are applied to the corresponding sensors to finally obtain a standardized sensor group.

[0037] For the measurement error model, first collect and prepare historical data, including historical environmental information, historical sensor initial parameters and historical sensor calibration parameters; when building the model, select an appropriate neural network architecture to train the measurement error model. The input layer of the neural network will receive environmental information and sensor initial parameters, while the output layer is the sensor calibration parameters, that is, the corrected sensor data; specifically, the neural network will learn the relationship between the input environmental information (such as temperature, humidity, etc.) and the sensor initial parameters, so as to predict the possible error of the sensor and give a calibration parameter; during the training process, using historical environmental data and historical sensor initial parameters, the neural network will learn how environmental changes affect the measurement error of the sensor. The goal of the model is to minimize the output calibration error, so the mean square error (MSE) is used as the loss function, and the weights of the neural network are optimized through the back propagation algorithm; through multiple rounds of training, the neural network gradually adjusts its internal parameters so that the predicted calibration parameters can correct the initial measurement error of the sensor as accurately as possible; once the model training is completed, the calibration parameters can be calculated by receiving the current environmental information and the current sensor initial parameters, and these calibrated parameters can be used for further data analysis or decision-making process.

[0038] After the sensor group calibration is completed, the system terminal will determine the data perception frequency based on the sensor's response speed and actual needs, and use the standard sensor group to monitor the material status in the environment in real time according to this frequency, and obtain a continuous material perception data stream. This data will be transmitted to the system terminal in real time for subsequent analysis and processing.

[0039] An AGV dynamic analysis model is obtained according to an adaptive algorithm, and a docking analysis is performed on the perception parameters based on the AGV dynamic analysis model to obtain applicable docking parameters.

[0040] In one embodiment, the system terminal establishes a dynamic analysis model of the AGV through an adaptive algorithm; the adaptive algorithm determines the model framework of the AGV dynamic analysis model based on the material docking database of the target AGV, and uses the material docking database to train the model framework, thereby constructing an AGV dynamic analysis model; once the dynamic analysis model is established, the system terminal performs docking analysis on the perception parameters (such as the relative position, height difference, depth, etc. of the material) based on the model, judges the docking situation between the material and the AGV through the learned knowledge, and calculates the docking parameters suitable for the current situation, that is, the applicable docking parameters, to ensure that the AGV can accurately dock with the material. These parameters include but are not limited to appropriate docking height, depth, platform position, etc., to ensure that the material can accurately dock with the AGV platform, thereby improving docking accuracy and production efficiency.

[0041] Furthermore, the step of obtaining the AGV dynamic analysis model according to the adaptive algorithm call includes:

[0042] Acquire the material docking database of the target AGV; acquire the AGV docking target, disassemble the AGV docking target, and determine the docking position, docking height, and docking depth; perform characteristic analysis on the docking position, docking height, and docking depth according to an adaptive algorithm, and select an AGV docking analysis model framework; train and optimize the material docking database based on the AGV docking analysis model framework to obtain the AGV dynamic analysis model.

[0043] Preferably, the system terminal collects and organizes historical data related to AGV material docking, which includes past docking conditions; these data can form a material docking database, which provides a basis for subsequent analysis; the information in the material docking database helps to understand the characteristics of different materials and the parameters required for AGV to dock with materials.

[0044] The system terminal selects a specific docking target, that is, the material that needs to be docked with the AGV, based on the information in the material docking database, and breaks down the docking process of the material into multiple key parameters, namely, docking position, docking height and docking depth; among them, the docking position refers to the spatial relationship between the material and the AGV platform; the docking height refers to the vertical distance between the material and the platform; the docking depth is the insertion depth of the material on the platform. These parameters are the basis for precise docking and determine whether the contact between the material and the AGV is accurate.

[0045] After determining the docking position, docking height and docking depth, the system terminal uses an adaptive algorithm to conduct a preliminary analysis of these perception data to find out the relationship between the docking parameters; for example, it analyzes how the height and depth affect the docking accuracy when the docking position of the material changes, or how the docking depth and position interact with each other at different docking heights to affect whether the material can be accurately docked with the AGV platform; the role of the adaptive algorithm is to analyze and select the most suitable AGV docking analysis model framework based on the current docking task requirements; specifically, the algorithm uses docking parameters such as docking position, docking height and docking depth as input, analyzes the relationship between these parameters in the current docking scenario, and selects a suitable analysis model based on this. model framework; for example, for a standard material docking task, the algorithm will evaluate the adaptability of different model frameworks in processing these parameters based on the mutual influence between the docking position, docking height and docking depth; if the docking task requires higher precision, a model that focuses on fine-tuning the docking depth and height may be selected, such as the Kalman filter model; if the docking task involves the coordination of multiple variables, a comprehensive analysis model framework that can simultaneously consider the position, depth and height may be selected, such as a multivariate regression model or a neural network module; by analyzing the dynamic relationship between these docking parameters, the algorithm can intelligently select the analysis model framework that best suits the current docking task, thereby ensuring the docking accuracy and success rate of the AGV and the material.

[0046] The system terminal is based on the selected AGV docking analysis model framework and uses the material docking database for training and optimization. By learning from historical data, the model will optimize parameters so that the analysis framework can handle future docking tasks more accurately. The result of the training is an optimized AGV dynamic analysis model, which can dynamically adjust the docking strategy based on real-time input data, thereby improving the accuracy and efficiency of AGV and material docking.

[0047] Furthermore, obtaining the AGV dynamic analysis model includes:

[0048] Based on the AGV docking analysis model framework, the material docking database is analyzed, trained and optimized until the preset analysis accuracy is achieved, and a docking position analysis model, a docking height analysis model and a docking depth analysis model are obtained; the docking position analysis model, the docking height analysis model and the docking depth analysis model are weightedly fused to obtain the AGV dynamic analysis model.

[0049] Optionally, the system terminal uses the data in the material docking database for training and optimization according to the selected AGV docking analysis model framework, such as the neural network model; the data in the database includes historical material docking conditions, docking positions, docking heights, docking depths and other information. During the training process, the model will analyze these data, find out the relationship between docking parameters, and optimize according to these relationships until the preset analysis accuracy is reached; specifically, the training process will independently analyze the docking position, docking height and docking depth, and construct three special analysis models, namely the docking position analysis model, the docking height analysis model and the docking depth analysis model, each model is deeply optimized for a specific docking parameter, and the model will continuously learn and adjust the historical data so that each analysis model can accurately predict the changes in the corresponding parameters under different docking conditions, and give corresponding corrections and optimization suggestions; during the training process of each analysis model, the system terminal uses the mean square error MSE to evaluate the accuracy of the model in the same way as mentioned above, and adjusts the parameters of the model through optimization algorithms such as gradient descent to ensure that it gradually reaches the preset analysis accuracy.

[0050] After obtaining independent docking position, docking height and docking depth analysis models, the system terminal merges the output results of the three models through weighted fusion technology. The purpose of weighted fusion is to comprehensively consider the mutual influence between the three parameters, so that the final model can simultaneously optimize the docking position, docking height and docking depth, and provide more accurate analysis results; specifically, the system terminal assigns different weights to each analysis model according to different tasks and environmental conditions; for example, in some cases, the accuracy of the docking position may be more important than the docking depth, then the weight of the docking position analysis model will be increased accordingly; weighted fusion can use linear weighted averaging, principal component analysis (PCA) or weighted regression methods to fuse the output results of each independent model to obtain the final AGV dynamic analysis model; the AGV dynamic analysis model is a comprehensive model, which can evaluate and adjust the docking position, docking height and docking depth in real time during the actual material docking process, thereby ensuring that the AGV can complete the docking task accurately and efficiently.

[0051] The multidimensional characteristic information of the target material is obtained, and the applicable docking parameters are optimized and analyzed based on the multidimensional characteristic information using an optimization control strategy to determine the target docking parameters.

[0052] In one embodiment, the system terminal first obtains multi-dimensional characteristic information of the target material, which includes the shape, size, weight, surface material, and stability of the material. Through sensors and other detection equipment, the system terminal can comprehensively collect various properties of the material. These characteristic information has a direct impact on the docking method and accuracy of the material. Subsequently, based on the collected material characteristic information, an optimization control strategy is used to optimize and analyze the applicable docking parameters. The optimization control strategy aims to obtain material docking parameters such as docking speed, platform extension length, etc. through global optimization according to the characteristics of the target material, and then use these material docking parameters to adjust the docking parameters (such as docking position, docking height, docking depth, etc.) to achieve the best docking effect. Finally, through the adjustment of the docking parameters, the system terminal determines the target docking parameters, which include the optimal docking position, docking height and docking depth, etc., to ensure accurate docking between the material and the AGV platform, improve production efficiency and reduce errors and failures.

[0053] Furthermore, the determining of the target docking parameters includes:

[0054] Based on the optimization control strategy, material docking data mining is performed to construct the material docking control space of the target AGV; the multidimensional characteristic information is used as a constraint parameter, and matched and divided with the material docking control space to obtain a material docking control memory library; a material docking fitness function is constructed, and the material docking fitness function is used to perform global optimization in the material docking control memory library to obtain material docking parameters; based on the material docking parameters, the applicable docking parameters are supplemented with optimization analysis to determine the target docking parameters.

[0055] Optionally, the system terminal uses an optimization control strategy to perform data mining on the material docking process, analyze the docking requirements of the materials and their interaction with the AGV platform; by collecting a large amount of historical data and counting the maximum and minimum values ​​of each possible variable, a material docking control space is constructed. This space describes the relationship between all possible variables and parameters in the material docking process. These variables may include the docking speed of the material, the extension length of the platform, the lifting height, etc. The construction of the material docking control space is based on a comprehensive analysis of the material docking task to ensure that various docking scenarios can be simulated and provide data support for subsequent optimization.

[0056] The system terminal uses the multi-dimensional characteristic information of the material (such as shape, size, weight, etc.) as constraint parameters and matches it with the variables in the material docking control space. These constraint parameters limit the feasibility range of the material docking process. For example, the weight and size of certain materials may limit the maximum length of the platform extension, or the shape of the material may affect the docking speed. By matching these characteristic information with the variables in the control space, the system terminal can divide a material docking control memory library, which stores the successful experience and parameter settings of different material docking tasks in the past as a basis for optimization and decision-making.

[0057] In order to ensure the success of the material docking task, the system terminal constructs a material docking fitness function, which is used to measure the pros and cons of different material docking parameters, so as to find a set of optimal material docking parameters, including the optimal docking speed, the optimal platform extension length, and the optimal lifting height; the material docking fitness function is as follows: F fit =w1*|v act -v tar ∣+w2*∣L act -L tar ∣+w3*∣h act -h tar ∣, where F fit is the adaptability of the docking parameters of this group of materials, v act is the docking speed of this group of material docking parameters, v tar is the expected docking speed, which is the average of all historical docking speeds that meet the constraint parameters, L act The platform extension length of the material docking parameters for this group, L tar is the expected platform extension length, which is the mean of all historical platform extension lengths that meet the constraint parameters, h act为 The lifting height of this group of material docking parameters, h tar is the predetermined lifting height, which is the average of all historical lifting heights that meet the constraint parameters. w1, w2, and w3 are the weights of docking speed, platform extension length, and lifting height, respectively. The sum of w1, w2, and w3 is 1. The system terminal calculates the fitness of each set of material docking parameters through this material docking fitness function, and performs global optimization based on the calculated fitness, searching for the optimal material docking parameters in the material docking control memory library.

[0058] After obtaining the optimal material docking parameters, the system terminal will perform additional optimization analysis on the applicable docking parameters based on these parameters; specifically, the system terminal will further adjust and optimize the applicable docking parameters for the current docking task according to the docking speed, platform extension length and lifting height; for the docking position, when the optimal docking speed is faster, the adjustment range of the docking position will be automatically reduced to ensure that the distance between the AGV platform and the material remains stable and avoid docking position errors caused by excessive speed. If the docking speed is slower, the tolerance range of the docking position can be slightly increased to improve efficiency; for the docking depth, if the platform extension length is shorter, the docking depth of the material may be insufficient and needs to be reduced. The docking depth requirement is to ensure that the material can be normally docked to the platform. If the platform is extended for a long time, the material may be over-inserted into the platform. At this time, it is necessary to increase the adjustment amount of the docking depth to ensure that the material can be accurately inserted into the appropriate position of the platform to avoid over-insertion or poor contact; for the docking height, if the lifting height is large, the tolerance range of the docking height will be automatically reduced to ensure the docking accuracy between the material and the platform. If the lifting height is small, the tolerance range of the docking height can be increased to improve efficiency; finally, the adjusted docking position, docking height and docking depth are summarized to form the target docking parameters to ensure that the material docking task can be successfully completed under various working conditions.

[0059] Furthermore, the obtaining of material docking parameters includes:

[0060] The material docking fitness function is used to select docking parameters and evaluate the fitness of the material docking control memory library to obtain multiple docking parameter fitnesses; based on the multiple docking parameter fitnesses, the material docking control memory library is iteratively optimized until a preset termination condition is reached to obtain the material docking parameters.

[0061] Optionally, the system terminal uses the defined material docking fitness function to evaluate different docking parameter groups in the material docking control memory library, measures the fitness of each docking parameter group in the docking task, and thus obtains the fitness value of each docking parameter group. These fitness values ​​reflect the pros and cons of the corresponding parameter groups in the actual docking task.

[0062] Once the system obtains the fitness values ​​of multiple docking parameters, it will perform iterative optimization of parameters based on these fitness values. The purpose of the iterative optimization process is to find a set of optimal parameters by continuously adjusting the docking parameters (such as docking speed, platform extension length, lifting height, etc.) so that the success rate, accuracy and efficiency of the material docking task can reach the best state; specifically, the system terminal will use optimization algorithms (such as particle swarm optimization, genetic algorithm or gradient descent, etc.) to adjust and optimize the docking parameters in the memory library according to the fitness evaluation results; in each round of iteration, the algorithm will select docking parameters with better performance for combination or mutation, generate a new docking parameter group, and perform fitness evaluation again. This process will be repeated for multiple rounds until the preset termination condition is reached, that is, the fitness value is greater than the fitness threshold or the maximum number of iterations is reached; finally, after multiple iterations and optimizations, the system terminal will obtain a set of optimal material docking parameters, which can provide the best docking effect in different material docking tasks.

[0063] In summary, the embodiments of the present application have at least the following technical effects:

[0064] First, a sensor group is installed on the target AGV, the sensor group includes a laser rangefinder sensor, an ultrasonic sensor, an infrared sensor and a visual sensor, and the sensor group is used to monitor and obtain the material perception data stream in real time, analyze and process the material perception data stream, and obtain the perception parameters of the target AGV and the target material, the perception parameters include relative position, height difference and depth; then, the AGV dynamic analysis model is obtained according to the adaptive algorithm, and the perception parameters are docked and analyzed based on the AGV dynamic analysis model to obtain applicable docking parameters; further, the multi-dimensional characteristic information of the target material is obtained, and the optimization control strategy is used to optimize and analyze the applicable docking parameters based on the multi-dimensional characteristic information to determine the target docking parameters; finally, the target AGV is controlled to perform adaptive docking control on the target material based on the target docking parameters. The technical problems of pushing and bumping materials caused by environmental changes, inconsistent material height or errors, and insufficient platform docking accuracy during the material docking process of AGV are solved. By combining the adaptive algorithm and the optimization control strategy, the docking parameters are dynamically adjusted according to the multi-dimensional characteristic information of the material, so as to achieve the technical effect of improving the docking accuracy of AGV materials, reducing the pushing and bumping materials, and improving production efficiency.

[0065] Embodiment 2 is based on the same inventive concept as the AGV adaptive material docking control method in the above embodiment. Figure 2As shown, the present application provides an AGV adaptive material docking control system, wherein the system includes: a sensor group installation module 11: installing a sensor group on the target AGV, the sensor group including a laser ranging sensor, an ultrasonic sensor, an infrared sensor and a visual sensor; an analysis and processing module 12: acquiring material perception data streams through real-time monitoring of the sensor group, analyzing and processing the material perception data streams, and obtaining perception parameters of the target AGV and the target material, the perception parameters including relative position, height difference and depth; a docking analysis module 13: acquiring an AGV dynamic analysis model according to an adaptive algorithm, performing docking analysis on the perception parameters based on the AGV dynamic analysis model, and obtaining applicable docking parameters; an optimization analysis module 14: acquiring multi-dimensional characteristic information of the target material, optimizing and analyzing the applicable docking parameters based on the multi-dimensional characteristic information using an optimization control strategy, and determining target docking parameters; a docking control module 15: controlling the target AGV to perform adaptive docking control on the target material based on the target docking parameters.

[0066] Furthermore, the analysis and processing module 12 is used to perform the following method:

[0067] Perform noise characteristic analysis on the material perception data stream to obtain perception data noise characteristic information; perform denoising on the material perception data stream based on the perception data noise characteristic information to obtain a standard material perception data stream; perform feature extraction and feature fusion on the standard material perception data stream to obtain a material fusion perception feature set; perform key parameter analysis on the material fusion perception feature set to determine the perception parameters of the target AGV and target material.

[0068] Furthermore, the analysis and processing module 12 is used to perform the following method:

[0069] According to the sensor group, a multi-sensor feature extraction algorithm set is determined; according to the multi-sensor feature extraction algorithm set, feature extraction is performed on the standard material perception data stream respectively to obtain a multi-sensor perception feature set; and a data fusion algorithm is used to align and fuse the multi-sensor perception feature set to obtain the material fusion perception feature set.

[0070] Furthermore, the analysis and processing module 12 is used to perform the following method:

[0071] Acquire application environment information of the target AGV; perform offset calibration on each sensor in the sensor group based on the application environment information to obtain a standard sensor group; set a data perception frequency, use the standard sensor group according to the data perception frequency, and monitor and acquire the material perception data stream in real time.

[0072] Furthermore, the docking analysis module 13 is used to perform the following method:

[0073] Acquire the material docking database of the target AGV; acquire the AGV docking target, disassemble the AGV docking target, and determine the docking position, docking height, and docking depth; perform characteristic analysis on the docking position, docking height, and docking depth according to an adaptive algorithm, and select an AGV docking analysis model framework; train and optimize the material docking database based on the AGV docking analysis model framework to obtain the AGV dynamic analysis model.

[0074] Furthermore, the docking analysis module 13 is used to perform the following method:

[0075] Based on the AGV docking analysis model framework, the material docking database is analyzed, trained and optimized until the preset analysis accuracy is achieved, and a docking position analysis model, a docking height analysis model and a docking depth analysis model are obtained; the docking position analysis model, the docking height analysis model and the docking depth analysis model are weightedly fused to obtain the AGV dynamic analysis model.

[0076] Furthermore, the optimization analysis module 14 is used to perform the following method:

[0077] Based on the optimization control strategy, material docking data mining is performed to construct the material docking control space of the target AGV; the multidimensional characteristic information is used as a constraint parameter, and matched and divided with the material docking control space to obtain a material docking control memory library; a material docking fitness function is constructed, and the material docking fitness function is used to perform global optimization in the material docking control memory library to obtain material docking parameters; based on the material docking parameters, the applicable docking parameters are supplemented with optimization analysis to determine the target docking parameters.

[0078] Furthermore, the optimization analysis module 14 is used to perform the following method:

[0079] The material docking fitness function is used to select docking parameters and evaluate the fitness of the material docking control memory library to obtain multiple docking parameter fitnesses; based on the multiple docking parameter fitnesses, the material docking control memory library is iteratively optimized until a preset termination condition is reached to obtain the material docking parameters.

[0080] Embodiment three, based on the same inventive concept as the AGV adaptive material docking control method in the aforementioned embodiment, the present application provides a storage medium, in which a computer program is stored, and the processor implements the following steps when executing the computer program: installing a sensor group on the target AGV, the sensor group including a laser ranging sensor, an ultrasonic sensor, an infrared sensor and a visual sensor; acquiring material perception data streams through real-time monitoring by the sensor group, analyzing and processing the material perception data streams, and obtaining perception parameters of the target AGV and the target material, the perception parameters including relative position, height difference and depth; acquiring an AGV dynamic analysis model according to an adaptive algorithm, performing docking analysis on the perception parameters based on the AGV dynamic analysis model, and obtaining applicable docking parameters; acquiring multi-dimensional characteristic information of the target material, optimizing and analyzing the applicable docking parameters based on the multi-dimensional characteristic information using an optimization control strategy, and determining target docking parameters; controlling the target AGV to perform adaptive docking control on the target material based on the target docking parameters.

[0081] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0082] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0083] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. AGV adaptive material docking control method, characterized in that: The method comprises: Installing a sensor group on the target AGV, the sensor group including a laser ranging sensor, an ultrasonic sensor, an infrared sensor and a visual sensor; Acquire material perception data streams through real-time monitoring of the sensor group, analyze and process the material perception data streams, and obtain perception parameters of the target AGV and the target material, wherein the perception parameters include relative position, height difference, and depth; Acquire an AGV dynamic analysis model according to an adaptive algorithm, perform docking analysis on the perception parameters based on the AGV dynamic analysis model, and obtain applicable docking parameters; Acquire multidimensional characteristic information of the target material, optimize and analyze the applicable docking parameters based on the multidimensional characteristic information using an optimization control strategy, and determine target docking parameters; The target AGV is controlled based on the target docking parameter to perform adaptive docking control on the target material.

2. The AGV adaptive material docking control method according to claim 1, characterized in that: The obtaining of the perception parameters of the target AGV and the target material includes: Performing noise characteristic analysis on the material perception data stream to obtain perception data noise characteristic information; Performing denoising on the material perception data stream based on the perception data noise characteristic information to obtain a standard material perception data stream; Performing feature extraction and feature fusion on the standard material perception data stream to obtain a material fusion perception feature set; A key parameter analysis is performed on the material fusion perception feature set to determine the perception parameters of the target AGV and the target material.

3. The AGV adaptive material docking control method according to claim 2, characterized in that: The obtaining of the material fusion perception feature set includes: Determining a multi-sensor feature extraction algorithm set according to the sensor group; According to the multi-sensor feature extraction algorithm set, feature extraction is performed on the standard material perception data stream respectively to obtain a multi-sensor perception feature set; The multi-sensor perception feature sets are aligned and fused using a data fusion algorithm to obtain the material fusion perception feature set.

4. The AGV adaptive material docking control method according to claim 1, characterized in that: The real-time monitoring and acquisition of material perception data stream by the sensor group includes: Obtaining application environment information of the target AGV; Performing offset calibration on each sensor in the sensor group based on the application environment information to obtain a standard sensor group; A data sensing frequency is set, and the standard sensor group is used according to the data sensing frequency to monitor and acquire the material sensing data stream in real time.

5. The AGV adaptive material docking control method according to claim 1, characterized in that: The method of obtaining the AGV dynamic analysis model according to the adaptive algorithm includes: Collect and obtain the material docking database of the target AGV; Acquire an AGV docking target, disassemble the AGV docking target, and determine a docking position, a docking height, and a docking depth; Performing characteristic analysis on the docking position, docking height and docking depth according to an adaptive algorithm, and selecting an AGV docking analysis model framework; The material docking database is trained and optimized based on the AGV docking analysis model framework to obtain the AGV dynamic analysis model.

6. The AGV adaptive material docking control method according to claim 5, characterized in that: The obtaining of the AGV dynamic analysis model comprises: Based on the AGV docking analysis model framework, the material docking database is analyzed, trained and optimized until the preset analysis accuracy is achieved, and a docking position analysis model, a docking height analysis model and a docking depth analysis model are obtained; The docking position analysis model, the docking height analysis model and the docking depth analysis model are weightedly fused to obtain the AGV dynamic analysis model.

7. The AGV adaptive material docking control method according to claim 6, characterized in that: Determining the target docking parameters includes: Perform material docking data mining based on the optimization control strategy to construct the material docking control space of the target AGV; The multi-dimensional characteristic information is used as a constraint parameter to match and divide the material docking control space to obtain a material docking control memory library; Constructing a material docking fitness function, and using the material docking fitness function to perform global optimization in the material docking control memory library to obtain material docking parameters; Based on the material docking parameters, the applicable docking parameters are supplemented and optimized and analyzed to determine the target docking parameters.

8. The AGV adaptive material docking control method according to claim 7, characterized in that: The obtaining of material docking parameters includes: Using the material docking fitness function to perform docking parameter selection and fitness evaluation on the material docking control memory library to obtain multiple docking parameter fitnesses; Based on the fitness of the multiple docking parameters, the material docking control memory library is iterated to optimize the parameters until a preset termination condition is reached to obtain the material docking parameters. 9.AGV adaptive material docking control system, characterized in that: For implementing the AGV adaptive material docking control method according to any one of claims 1 to 8, the system comprises: Sensor group installation module: install a sensor group on the target AGV, the sensor group including a laser ranging sensor, an ultrasonic sensor, an infrared sensor and a visual sensor; Analysis and processing module: acquires material perception data stream through real-time monitoring of the sensor group, analyzes and processes the material perception data stream, and obtains perception parameters of the target AGV and the target material, wherein the perception parameters include relative position, height difference and depth; Docking analysis module: obtaining an AGV dynamic analysis model according to an adaptive algorithm, performing docking analysis on the perception parameters based on the AGV dynamic analysis model, and obtaining applicable docking parameters; Optimization analysis module: obtaining multi-dimensional characteristic information of the target material, optimizing and analyzing the applicable docking parameters based on the multi-dimensional characteristic information using an optimization control strategy, and determining target docking parameters; Docking control module: controls the target AGV to perform adaptive docking control on the target material based on the target docking parameters.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the AGV adaptive material docking control method as described in any one of claims 1 to 8 is implemented.

Images