A crane adaptive control method and control system

Abstract

The present application relates to a crane adaptive control method and control system thereof, and relates to the field of cranes. The control method includes: using artificial intelligence technology based on deep learning to collaboratively analyze the load data and environmental data of the crane hook, respectively extracting the time series dynamic change characteristics of the load, wind direction and wind speed, and interactively aggregating the fine-grained time series characteristics of the wind direction and wind speed to obtain the global time series characteristic representation of wind force that integrates multimodal information, and then based on the main component matching fusion characteristics of the load and wind force, realize adaptive control of the crane moving speed. In this way, it is possible to dynamically adapt to wind changes and reasonably control the moving speed of the crane, thereby improving the working stability and safety of the crane in complex environments.

Description

Technical Field

[0001] The present application relates to the technical field of cranes, and in particular, to a crane adaptive control method and a control system thereof. Background Art

[0002] With the rapid development of industrial automation, cranes, as important material handling equipment, play a vital role in various production environments. However, traditional cranes often rely on the operator's experience and judgment during operation, and it is difficult to accurately adapt to the complex and changing working environment, especially under the influence of external factors such as load changes and environmental impacts, its operating efficiency and safety face many challenges.

[0003] Cranes have very different stability and safety requirements when carrying loads of different weights. Overloading may damage the crane structure or even cause safety accidents, while underloading may cause material damage or inefficiency due to improper operation. In addition, for cranes operating outdoors or in open environments, wind is an external factor that cannot be ignored. Wind not only affects the stability of the crane, but also increases the difficulty of operation and safety hazards. Especially in windy weather, if the crane cannot adjust its moving speed and posture in time, it is very easy to overturn or collide.

[0004] However, most existing cranes are controlled by fixed parameters or preset programs, lacking the ability to perceive and adaptively adjust complex environmental factors in real time, and cannot flexibly respond to complex changes in actual operations. Therefore, in order to improve the operational safety of cranes, a crane adaptive control method and its control system are expected. Summary of the invention

[0005] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] In a first aspect, the present application provides a crane adaptive control method, the method comprising:

[0007] Obtaining a time queue of load values ​​collected by a load sensor, and a time queue of wind direction and a time queue of wind speed collected by an environmental sensor;

[0008] Transmitting the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed to an adaptive control center through a wireless communication module;

[0009] In the adaptive control center, time series feature extraction is performed on the time queue of the load value, the time queue of the wind direction, and the time queue of the wind speed to obtain a load time series associated implicit feature vector, a wind direction time series associated implicit feature vector, and a wind speed time series associated implicit feature vector;

[0010] Inputting the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector into a local fine-grained sequence interaction response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interaction global representation vector;

[0011] Based on the principal component matching fusion feature between the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector, a control instruction for the moving speed is determined.

[0012] Optionally, time series feature extraction is performed on the time queue of the load value, the time queue of the wind direction, and the time queue of the wind speed, respectively, to obtain implicit feature vectors associated with the load time series, implicit feature vectors associated with the wind direction, and implicit feature vectors associated with the wind speed time series, including: inputting the time queue of the load value, the time queue of the wind direction, and the time queue of the wind speed into a sequence encoder based on a Bi-LSTM model, respectively, to obtain implicit feature vectors associated with the load time series, implicit feature vectors associated with the wind direction, and implicit feature vectors associated with the wind speed time series.

[0013] Optionally, the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector are input into a local fine-grained sequence interactive response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interactive global representation vector, including: performing feature decoupling on the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector respectively to obtain a sequence of wind direction time series local representation vectors and a sequence of wind speed time series local representation vectors; calculating the wind direction time series local representation vector sequence and the wind speed time series local representation vector sequence of each group of corresponding wind direction time series local representation vectors and wind speed time series local representation vector sequence. The local semantic association interaction vectors between the wind power local temporal semantic association interaction vectors are obtained to obtain a sequence of wind power local temporal semantic association interaction vectors; the fine-grained attention weights of each wind power local temporal semantic association interaction vector in the sequence of wind power local temporal semantic association interaction vectors are calculated to obtain a sequence of fine-grained attention weights; the sequence of fine-grained attention weights is input into a random deactivation module to obtain a sequence of fine-grained attention weights after deactivation; the sequence of fine-grained attention weights after deactivation is used as a weight distribution, and the weighted sum of the sequence of wind power local temporal semantic association interaction vectors is calculated to obtain the global representation vector of the wind power temporal decoupling and enhanced interaction.

[0014] Optionally, the local semantic association interaction vector between each group of corresponding wind direction time series local representation vectors and wind speed time series local representation vectors in the sequence of the wind direction time series local representation vectors and the sequence of the wind speed time series local representation vectors is calculated to obtain a sequence of wind local time series semantic association interaction vectors, including: calculating the position point addition, position point subtraction and position point multiplication between the wind direction time series local representation vector and the corresponding wind speed time series local representation vector to obtain a first fusion result, a second fusion result and a third fusion result; after cascading the first fusion result, the second fusion result and the third fusion result into a wind multi-scale fusion feature vector, one-dimensional convolution processing and maximum pooling processing are performed on them to obtain the wind local time series semantic association interaction vector.

[0015] Optionally, the fine-grained attention weight of each wind local temporal semantic association interaction vector in the sequence of wind local temporal semantic association interaction vectors is calculated to obtain a sequence of fine-grained attention weights, including: multiplying the wind local temporal semantic association interaction vector by a weight coefficient matrix, adding the result to the bias vector and then passing the result through a tanh function to obtain a wind local temporal semantic association interaction encoding feature vector; using an attention score reference vector to multiply the wind local temporal semantic association interaction encoding feature vector to obtain the fine-grained attention weight.

[0016] Optionally, based on the principal component matching fusion features between the load timing-associated implicit feature vector and the wind timing decoupling and enhanced interaction global representation vector, a control instruction for the moving speed is determined, including: inputting the load timing-associated implicit feature vector and the wind timing decoupling and enhanced interaction global representation vector into a significant fusion network based on feature principal component query matching to obtain a load-wind principal component collaborative interaction fusion representation vector; inputting the load-wind principal component collaborative interaction fusion representation vector into a decoder-based moving speed controller to obtain the control instruction, wherein the control instruction is used to represent a recommended moving speed decoding value.

[0017] Optionally, the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector are input into a significant fusion network based on feature principal component query matching to obtain a load-wind principal component collaborative interaction fusion representation vector, including: standardizing the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector to obtain a standardized load time series associated implicit feature vector and a standardized wind time series decoupling enhanced interaction global representation vector; respectively calculating the sample covariance matrices of the standardized load time series associated implicit feature vector and the standardized wind time series decoupling enhanced interaction global representation vector to obtain a load sample covariance matrix and a wind sample covariance matrix; performing matrix decomposition-based feature vector extraction on the load sample covariance matrix and the wind sample covariance matrix to obtain a set of load time series associated principal component feature vectors and a set of wind time series interaction principal component feature vectors; and inputting the set of load time series associated principal component feature vectors and the set of wind time series interaction principal component feature vectors into a significant fusion network based on feature principal component query matching to obtain a load-wind principal component collaborative interaction fusion representation vector. The method comprises the following steps: inputting a maximum approximate query matching network to obtain a set of best matching pairs of the load time series associated principal component feature vector and the wind time series interactive principal component feature vector; inputting each best matching pair in the set of best matching pairs of the load time series associated principal component feature vector and the wind time series interactive principal component feature vector into a fine-grained semantic interaction joint module to obtain a set of load-wind principal component feature vectors; cascading the set of load-wind principal component feature vectors to obtain the load-wind principal component collaborative interaction fusion representation vector; wherein, inputting each best matching pair in the set of best matching pairs of the load time series associated principal component feature vector and the wind time series interactive principal component feature vector into a fine-grained semantic interaction joint module to obtain a set of load-wind principal component feature vectors, including: performing multi-dimensional feature fusion on the best matching pairs, and after cascading the multi-dimensional feature fusion results into a load-wind principal component multi-scale fusion feature vector, performing one-dimensional convolution processing and maximum value pooling processing on the result to obtain the load-wind principal component feature vector.

[0018] Optionally, the load sample covariance matrix and the wind sample covariance matrix are subjected to eigenvector extraction based on matrix decomposition to obtain a set of load time series associated principal component eigenvectors and a set of wind time series interactive principal component eigenvectors, including: calculating the principal component eigenvalues ​​and principal component direction eigenvectors of the wind sample covariance matrix, so that the wind sample covariance matrix multiplied by the principal component direction eigenvector is equal to the principal component eigenvalue multiplied by the principal component direction eigenvector, so as to obtain a set of wind time series interactive principal component eigenvalues ​​and A set of wind time series interactive principal component directional eigenvectors; using the set of wind time series interactive principal component eigenvalues ​​as diagonal elements to construct a wind time series interactive principal component eigenvalue diagonal matrix; calculating the product of each wind time series interactive principal component directional eigenvector in the set of wind time series interactive principal component directional eigenvectors and the wind time series interactive principal component eigenvalue diagonal matrix, and then multiplying the product by the transposed vector of the wind time series interactive principal component directional eigenvector to obtain the set of wind time series interactive principal component eigenvectors.

[0019] Optionally, the set of load time series associated principal component feature vectors and the set of wind time series interaction principal component feature vectors are input into a maximum approximate query matching network to obtain a set of best matching pairs of load time series associated principal component feature vectors and wind time series interaction principal component feature vectors, including: calculating the cosine similarity between the i-th load time series associated principal component feature vector in the set of load time series associated principal component feature vectors and each wind time series interaction principal component feature vector in the set of wind time series interaction principal component feature vectors to obtain a set of cosine similarities; and forming the best matching pair with the i-th load time series associated principal component feature vector corresponding to the maximum value in the set of cosine similarities.

[0020] In a second aspect, the present application provides a crane adaptive control system, the system comprising:

[0021] A load value, wind direction and wind speed acquisition module, used to acquire a time queue of load values ​​collected by a load sensor, and a time queue of wind direction and a time queue of wind speed collected by an environmental sensor;

[0022] A wireless communication module, used for transmitting the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed to an adaptive control center through the wireless communication module;

[0023] A time series feature extraction module is used to extract time series features of the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed in the adaptive control center to obtain a load time series associated implicit feature vector, a wind direction time series associated implicit feature vector and a wind speed time series associated implicit feature vector;

[0024] A local fine-grained sequence interaction response aggregation module is used to input the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector into a local fine-grained sequence interaction response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interaction global representation vector;

[0025] The control instruction determination module for the moving speed is used to determine the control instruction for the moving speed based on the principal component matching fusion feature between the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector.

[0026] The above technical solution is adopted to collaboratively analyze the load data and environmental data of the crane hook by using artificial intelligence technology based on deep learning, and the time series dynamic change characteristics of load, wind direction and wind speed are extracted respectively. The wind direction and wind speed are interactively aggregated with fine-grained time series characteristics to obtain the global time series characteristic representation of wind force that integrates multimodal information, and then the adaptive control of the crane moving speed is realized based on the main component matching fusion characteristics of load and wind force. In this way, it is possible to dynamically adapt to wind changes and reasonably control the moving speed of the crane, thereby improving the working stability and safety of the crane in complex environments.

[0027] Other features and advantages of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0029] Figure 1 The present invention is a flowchart of a crane adaptive control method according to an exemplary embodiment.

[0030] Figure 2 It is a block diagram of a crane adaptive control system according to an exemplary embodiment.

[0031] Figure 3 It is a block diagram of an electronic device according to an exemplary embodiment.

[0032] Figure 4It is an application scenario diagram of a crane adaptive control method according to an exemplary embodiment. DETAILED DESCRIPTION

[0033] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments described herein. Instead, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0034] It should be understood that the various steps described in the method implementation of the present application can be performed in different orders and / or performed in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present application is not limited in this respect.

[0035] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0036] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0037] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0038] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0039] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.

[0040] Figure 1 is a flow chart of a crane adaptive control method according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0041] Step S101, obtaining a time queue of load values ​​collected by a load sensor, and a time queue of wind direction and a time queue of wind speed collected by an environmental sensor;

[0042] Step S102, transmitting the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed to an adaptive control center through a wireless communication module;

[0043] Step S103, in the adaptive control center, respectively extracting time series features from the time queue of the load value, the time queue of the wind direction, and the time queue of the wind speed to obtain a load time series associated implicit feature vector, a wind direction time series associated implicit feature vector, and a wind speed time series associated implicit feature vector;

[0044] Step S104: input the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector into a local fine-grained sequence interaction response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interaction global representation vector;

[0045] Step S105: Determine a control instruction for the moving speed based on the principal component matching fusion feature between the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector.

[0046] In response to the above technical problems, the technical concept of this application is to use artificial intelligence technology based on deep learning to collaboratively analyze the load data and environmental data of the crane hook, extract the time-series dynamic change characteristics of the load, wind direction and wind speed respectively, and interactively aggregate the fine-grained time-series characteristics of the wind direction and wind speed to obtain the global time-series characteristic representation of wind power that integrates multimodal information, and then realize adaptive control of the crane's moving speed based on the main component matching fusion characteristics of the load and wind power. In this way, it is possible to dynamically adapt to wind changes and reasonably control the crane's moving speed, thereby improving the working stability and safety of the crane in complex environments.

[0047] First, obtain the time queue of the load value collected by the load sensor, as well as the time queue of the wind direction and the time queue of the wind speed collected by the environmental sensor. It should be understood that the load value is one of the core parameters in the operation process of the crane, which is directly related to the stability and safety of the crane. Wind direction and wind speed are important environmental factors affecting the operation of the crane, which directly affect the stability and operation difficulty of the crane. Therefore, in the technical solution of the present application, by collecting the load data of the crane and the wind direction and wind speed data in the environment, the actual working status and environmental conditions of the crane can be grasped in real time, and this is used as the data support for the adaptive control of the crane.

[0048] Then, considering the limitation of local computing resources, the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed are further transmitted to the adaptive control center through the wireless communication module, so as to utilize the powerful computing resources of the adaptive control center for centralized data processing and analysis. In an embodiment of the present application, wireless communication technologies such as Wi-Fi, Zigbee or 5G can be used to realize real-time transmission of data to ensure the integrity and timeliness of the data.

[0049] Furthermore, considering that the load, wind direction and wind speed data all present complex dynamic change characteristics in the time dimension. Therefore, in order to effectively capture the implicit features and change rules in these time series data, this application uses the Bi-LSTM (bidirectional long short-term memory network) model as a sequence encoder to perform time series encoding on the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed, respectively, so as to utilize the bidirectional memory characteristics and gating mechanism of Bi-LSTM to fully explore the long-term dependencies and dynamic change trends in the time queue data, thereby obtaining the implicit feature vector associated with the load time series, the implicit feature vector associated with the wind direction time series and the implicit feature vector associated with the wind speed time series.

[0050] In one embodiment of the present application, time series feature extraction is performed on the time queue of the load value, the time queue of the wind direction, and the time queue of the wind speed, respectively, to obtain load time series associated implicit feature vectors, wind direction time series associated implicit feature vectors, and wind speed time series associated implicit feature vectors, including: inputting the time queue of the load value, the time queue of the wind direction, and the time queue of the wind speed into a sequence encoder based on a Bi-LSTM model, respectively, to obtain the load time series associated implicit feature vectors, the wind direction time series associated implicit feature vectors, and the wind speed time series associated implicit feature vectors.

[0051] It should be understood that wind direction determines the direction of wind force, while wind speed reflects the magnitude of wind force. The two together constitute the main external conditions that need to be considered when a crane operates in a wind field. Therefore, in order to comprehensively consider the interaction between wind direction and wind speed, this application introduces a local fine-grained sequence interaction response aggregation network based on feature decoupling to perform fine-grained feature interaction response analysis on the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector. Specifically, the network first decouples the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector respectively, decomposing the two into two sets of local representation vector sequences to enhance the sensitivity to local detail features. Then, the local semantic association calculation is performed on the sequence of the two sets of local representation vectors after decoupling to generate a local time series semantic association interaction vector sequence, and then the strength of the representation feature interaction response is further quantified to generate an attention weight sequence, so that the model can focus on the most important feature interactions, thereby improving the quality of feature representation and prediction accuracy. At the same time, in order to prevent overfitting, the generated attention weight sequence is further subjected to random inactivation processing to enhance the generalization ability of the model. Finally, the attention weight sequence after random inactivation processing is used as the weight distribution to perform weighted summation on the local time series semantic association interaction vector sequence to synthesize the local interaction information of the wind direction time series association implicit feature vector and the wind speed time series association implicit feature vector, and obtain the global wind power time series decoupling enhanced interaction global representation vector, thereby achieving a comprehensive characterization of wind power characteristics.

[0052] In one embodiment of the present application, the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector are input into a local fine-grained sequence interactive response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interactive global representation vector, including: performing feature decoupling on the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector respectively to obtain a sequence of wind direction time series local representation vectors and a sequence of wind speed time series local representation vectors; calculating the wind direction time series local representation vector and the wind speed time series local representation vector in the sequence of the wind direction time series local representation vector and the wind speed time series local representation vector for each group of corresponding wind direction time series local representation vectors and wind speed time series local representation vectors; The local semantic association interaction vectors between the wind power vectors are obtained to obtain a sequence of wind power local temporal semantic association interaction vectors; the fine-grained attention weights of each wind power local temporal semantic association interaction vector in the sequence of wind power local temporal semantic association interaction vectors are calculated to obtain a sequence of fine-grained attention weights; the sequence of fine-grained attention weights is input into a random deactivation module to obtain a sequence of fine-grained attention weights after deactivation; the sequence of fine-grained attention weights after deactivation is used as a weight distribution, and the weighted sum of the sequence of wind power local temporal semantic association interaction vectors is calculated to obtain the wind power temporal decoupling enhanced interaction global representation vector.

[0053] Furthermore, in one embodiment of the present application, the local semantic association interaction vector between each group of corresponding wind direction time series local representation vectors and wind speed time series local representation vectors in the sequence of the wind direction time series local representation vectors and the sequence of the wind speed time series local representation vectors is calculated to obtain a sequence of wind local time series semantic association interaction vectors, including: calculating the position point addition, position point subtraction and position point multiplication between the wind direction time series local representation vector and the corresponding wind speed time series local representation vector to obtain a first fusion result, a second fusion result and a third fusion result; after cascading the first fusion result, the second fusion result and the third fusion result into a wind multi-scale fusion feature vector, one-dimensional convolution processing and maximum pooling processing are performed on them to obtain the wind local time series semantic association interaction vector.

[0054] Furthermore, in one embodiment of the present application, the fine-grained attention weights of each wind local temporal semantic association interaction vector in the sequence of the wind local temporal semantic association interaction vectors are calculated to obtain a sequence of fine-grained attention weights, including: multiplying the wind local temporal semantic association interaction vector by a weight coefficient matrix, adding the vector to the bias vector and then passing the result through a tanh function to obtain a wind local temporal semantic association interaction encoding feature vector; using an attention score reference vector to multiply the wind local temporal semantic association interaction encoding feature vector to obtain the fine-grained attention weight.

[0055] Specifically, the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector are processed by the following interactive response analysis formula to obtain the wind power time series decoupling enhanced interaction global representation vector:

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062] in, represents the wind direction time series associated implicit eigenvector, represents the wind speed time series associated implicit feature vector, Indicates decoupling processing, , and Respectively represent the first, second, and third in the sequence of the local representation vector of the wind direction time series and A local representation vector of wind direction time series, The value of is the number of eigenvectors in the sequence of the local representation vector of the wind direction time series, , , and respectively represent the first, second, and third vectors in the sequence of the local representation vector of the wind speed time series. and A local representation vector of wind speed time series, Indicates cascade operation, represents the maximum pooling operation, represents a one-dimensional convolution operation, represents dot product, Indicates point addition, Indicates point reduction, The first in the sequence of the wind local temporal semantic association interaction vector Wind local temporal semantic association interaction vector, is the weight coefficient matrix, is the bias vector, is the hyperbolic tangent function, is the attention score reference vector, Indicates fine-grained attention weights, represents random dropout processing, Indicates The fine-grained attention weights after dropout processing, represents the global representation vector of the wind time series decoupling and enhancement interaction, where n is a positive integer greater than or equal to 2.

[0063] Then, in order to comprehensively consider the influence of load factors and wind factors on the stability of the crane, the present application further introduces a significant fusion network based on feature principal component query matching to match and analyze the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector and information aggregation to obtain a comprehensive representation of the crane operation state. Specifically, the significant fusion network first performs a standardization operation on the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector to ensure that the data is evenly distributed on the scale of zero mean and has unit variance, thereby eliminating the effects of different dimensions and magnitudes. Subsequently, the standardized load characteristics and wind characteristics are subjected to feature sparse processing by the principal component analysis algorithm, and the principal component characteristics are extracted to form a set of load time series associated principal component feature vectors and a set of wind time series interaction principal component feature vectors, respectively. Then, based on the maximum approximate query matching principle between feature vectors, the best matching pair is found by calculating and comparing the similarity between the two principal component feature vector sets. Next, these best matching pairs are subjected to fine-grained semantic joint operations, and the feature fusion results of each best matching pair are cascaded to aggregate the core component association fusion information of load and wind characteristics, and generate load-wind principal component collaborative interaction fusion representation vectors. In this way, the dynamic interaction between load and wind can be effectively captured, the feature expression ability can be further enhanced, and a more accurate feature reference can be provided for the adaptive control of the crane.

[0064] In one embodiment of the present application, based on the principal component matching fusion features between the load timing-associated implicit feature vector and the wind timing decoupling and enhanced interaction global representation vector, a control instruction for the moving speed is determined, including: inputting the load timing-associated implicit feature vector and the wind timing decoupling and enhanced interaction global representation vector into a significant fusion network based on feature principal component query matching to obtain a load-wind principal component collaborative interaction fusion representation vector; inputting the load-wind principal component collaborative interaction fusion representation vector into a decoder-based moving speed controller to obtain the control instruction, wherein the control instruction is used to represent a recommended moving speed decoding value.

[0065] Further, in one embodiment of the present application, the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector are input into a significant fusion network based on feature principal component query matching to obtain a load-wind principal component collaborative interaction fusion representation vector, including:

[0066] The load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector are standardized to obtain a standardized load time series associated implicit feature vector and a standardized wind time series decoupling enhanced interaction global representation vector;

[0067] Respectively calculating the sample covariance matrices of the standardized load time series associated implicit eigenvector and the standardized wind time series decoupled enhanced interaction global representation vector to obtain a load sample covariance matrix and a wind sample covariance matrix;

[0068] Performing matrix decomposition-based eigenvector extraction on the load sample covariance matrix and the wind sample covariance matrix to obtain a set of load time series associated principal component eigenvectors and a set of wind time series interactive principal component eigenvectors;

[0069] Inputting the set of the load time series associated principal component feature vectors and the set of the wind time series interactive principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of the load time series associated principal component feature vectors and the wind time series interactive principal component feature vectors;

[0070] Input each best matching pair in the set of best matching pairs of the load time series association principal component feature vector and the wind time series interaction principal component feature vector into a fine-grained semantic interaction joint module to obtain a set of load-wind principal component feature vectors;

[0071] The set of the load-wind principal component feature vectors is cascaded to obtain the load-wind principal component collaborative interaction fusion representation vector; wherein each best matching pair in the set of the best matching pairs of the load time series association principal component feature vector and the wind time series interaction principal component feature vector is input into a fine-grained semantic interaction joint module to obtain a set of load-wind principal component feature vectors, including: performing multi-dimensional feature fusion on the best matching pairs, and cascading the multi-dimensional feature fusion results into a load-wind principal component multi-scale fusion feature vector, and then performing one-dimensional convolution processing and maximum pooling processing on it to obtain the load-wind principal component feature vector.

[0072] Further, in one embodiment of the present application, the load sample covariance matrix and the wind sample covariance matrix are subjected to matrix decomposition-based eigenvector extraction to obtain a set of load time series associated principal component eigenvectors and a set of wind time series interactive principal component eigenvectors, including: calculating the principal component eigenvalues ​​and principal component direction eigenvectors of the wind sample covariance matrix, so that the wind sample covariance matrix multiplied by the principal component direction eigenvector is equal to the principal component eigenvalue multiplied by the principal component direction eigenvector, so as to obtain the wind time series interactive principal component eigenvector. A set of eigenvalues ​​and a set of wind time series interaction principal component direction eigenvectors; using the set of wind time series interaction principal component eigenvalues ​​as diagonal elements to construct a wind time series interaction principal component eigenvalue diagonal matrix; calculating the product of each wind time series interaction principal component direction eigenvector in the set of wind time series interaction principal component direction eigenvectors and the wind time series interaction principal component eigenvalue diagonal matrix, and then multiplying the product by the transposed vector of the wind time series interaction principal component direction eigenvector to obtain the set of wind time series interaction principal component eigenvectors.

[0073] Furthermore, in one embodiment of the present application, the set of the load timing associated principal component feature vectors and the set of the wind timing interaction principal component feature vectors are input into a maximum approximation query matching network to obtain a set of best matching pairs of the load timing associated principal component feature vectors and the wind timing interaction principal component feature vectors, including: calculating the cosine similarity between the i-th load timing associated principal component feature vector in the set of the load timing associated principal component feature vectors and each wind timing interaction principal component feature vector in the set of the wind timing interaction principal component feature vectors to obtain a set of cosine similarities; the wind timing interaction principal component feature vector corresponding to the maximum value in the set of cosine similarities and the i-th load timing associated principal component feature vector form the best matching pair.

[0074] Specifically, the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector are fused using the following principal component query matching formula to obtain the load-wind principal component collaborative interaction fusion representation vector, wherein the principal component query matching formula is:

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088] in, represents the load timing associated implicit feature vector, and represent the mean and standard deviation of the load timing associated implicit feature vector, respectively, represents the normalized load timing-related implicit eigenvector, represents the global representation vector of the wind power time series decoupling and enhancement interaction, and They represent the mean and standard deviation of the global representation vector of the wind time series decoupling and enhancement interaction, represents the global representation vector of the standardized wind time series decoupling and enhanced interaction, The value of is the length of the implicit characteristic vector associated with the load timing, The value of is the length of the global representation vector of the wind time series decoupling and enhancement interaction, represents the load sample covariance matrix, represents the wind sample covariance matrix, represents the set of eigenvectors of the principal component directions of the load time series, , and They represent the first, second and third eigenvectors of the direction of the principal component of the load time series. The load time series associated principal component directional eigenvectors, , and They represent the first, second and third eigenvalues ​​of the load time series correlation principal component. The load time series are associated with the principal component eigenvalues, The value of is the number of eigenvalues ​​of the principal component associated with the load time series, represents the set of eigenvectors of the interactive principal components of the wind time series, , and They represent the first, second and third eigenvectors of the wind time series interaction principal component direction. The wind time series interactive principal component directional eigenvectors, , and They represent the first, second and third eigenvalues ​​of the interactive principal component of the wind time series. The interactive principal component eigenvalues ​​of wind time series, is the diag function, which is used to construct a diagonal matrix. and They represent the diagonal matrix of the principal component eigenvalues ​​of the load time series association and the diagonal matrix of the principal component eigenvalues ​​of the wind time series interaction, and represent respectively the set of the load time series associated principal component eigenvectors and the set of the wind time series interactive principal component eigenvectors, Indicates The load time series associated principal component eigenvectors, and Respectively represent the first and The interactive principal component eigenvectors of wind time series, represents the inner product, represents the magnitude of the eigenvector, Indicates the index corresponding to the maximum value. Indicates that The index of the wind power time series interaction principal component eigenvector that constitutes the best matching pair of load time series associated principal component eigenvectors, Indicates cascade operation, represents the maximum pooling operation, represents a one-dimensional convolution operation, represents dot product, Indicates point addition, Indicates point reduction, The first element in the set of load-wind principal component eigenvectors load-wind principal component eigenvectors, A representation vector representing the load-wind principal component collaborative interaction fusion.

[0089] Next, the load-wind principal component collaborative interaction fusion representation vector is input into a decoder-based mobile speed controller to obtain a control instruction, and the control instruction is used to represent the recommended mobile speed decoding value. In the technical solution of the present application, the decoder adopts a multi-layer perceptron structure, and performs feature learning and decoding regression on the load-wind principal component collaborative interaction fusion representation vector through multi-layer nonlinear transformation to map the load-wind principal component collaborative interaction fusion representation vector to the mobile speed numerical space, and generate a mobile speed decoding value adapted to the current load state and environmental conditions, thereby realizing dynamic adaptive control of the crane.

[0090] In a preferred example, considering that the load timing association implicit feature vector and the wind timing decoupling enhanced interaction global representation vector respectively represent the wind direction and wind speed timing decoupling enhanced interaction characteristics and the load local timing association characteristics, therefore, when performing significant fusion based on feature principal component query matching, due to the difference in timing association feature distribution caused by the difference in source timing distribution and the difference in timing feature encoding method, the load-wind principal component collaborative interaction fusion representation vector will also have a timing feature diversified set expression distribution. Therefore, it is expected to improve the balance between the regression mapping accuracy and completeness of the load-wind principal component collaborative interaction fusion representation vector input when the mobile speed controller based on the decoder performs decoding regression, thereby improving the accuracy of the obtained control instructions.

[0091] Based on this, in the preferred example, inputting the load-wind principal component collaborative interaction fusion representation vector into a decoder-based mobile speed controller to obtain a control instruction specifically includes:

[0092] Determine the maximum eigenvalue of the load-wind principal component collaborative interaction fusion and the minimum eigenvalue of the load-wind principal component collaborative interaction fusion of the load-wind principal component collaborative interaction fusion representation vector, and calculate the load-wind principal component collaborative interaction fusion mean and load-wind principal component collaborative interaction fusion standard deviation of the feature set of the load-wind principal component collaborative interaction fusion representation vector;

[0093] Calculating the quotient of the load-wind principal component collaborative interaction fusion mean and the load-wind principal component collaborative interaction fusion standard deviation to obtain a load-wind principal component collaborative interaction fusion statistical normalization value;

[0094] Calculate the inverse of each eigenvalue of the load-wind principal component collaborative interaction fusion representation vector, multiply it by the difference between the maximum eigenvalue of the load-wind principal component collaborative interaction fusion and the minimum eigenvalue of the load-wind principal component collaborative interaction fusion, and then subtract it from the load-wind principal component collaborative interaction fusion statistical normalization value to obtain a load-wind principal component collaborative interaction fusion distribution approximation vector;

[0095] Calculate an exponential function with a natural constant as the base, where each eigenvalue of the load-wind principal component collaborative interaction fusion distribution approximation vector is an exponential function, and perform point addition with the load-wind principal component collaborative interaction fusion statistical normalization value to obtain a load-wind principal component collaborative interaction fusion probability approximation vector;

[0096] Calculate the base-two logarithm of the absolute value of each eigenvalue of the load-wind principal component collaborative interaction fusion probability approximation vector to obtain an optimized load-wind principal component collaborative interaction fusion representation vector; and input the optimized load-wind principal component collaborative interaction fusion representation vector into a decoder-based mobile speed controller to obtain a control instruction.

[0097] The load-wind principal component collaborative interaction fusion representation vector is denoted as The optimization expression is:

[0098]

[0099] , , and They are the feature sets of the load-wind principal component collaborative interaction fusion representation vector The mean, standard deviation, minimum and maximum values ​​of is the logarithm to base 2, represents the calculation of the reciprocal of each eigenvalue of the load-wind principal component collaborative interaction fusion representation vector, is the optimized load-wind principal component collaborative interaction fusion representation vector, Indicates adding by position point. It is the point multiplication by position. It is subtracted by position point.

[0100] That is, by performing random statistical normalization on the diversified feature set of the load-wind principal component collaborative interaction fusion representation vector, the confidence space constructed based on the overall eigenvalue of the load-wind principal component collaborative interaction fusion representation vector is approximated by a standardized continuous probability density distribution relative to the response hypothesis test of each eigenvalue of the load-wind principal component collaborative interaction fusion representation vector, thereby establishing the target reachability of the discrete feature distribution of the load-wind principal component collaborative interaction fusion representation vector to a unified regression target, so as to achieve a balanced executable performance between mapping accuracy and mapping completeness in the decoding regression process based on the discrete feature distribution of the load-wind principal component collaborative interaction fusion representation vector, and improve the accuracy of the control instructions obtained by the mobile speed controller based on the decoder input of the load-wind principal component collaborative interaction fusion representation vector.

[0101] In summary, the above scheme adopts the artificial intelligence technology based on deep learning to collaboratively analyze the load data and environmental data of the crane hook, extract the time series dynamic change characteristics of load, wind direction and wind speed respectively, and obtain the global time series feature representation of wind force that integrates multimodal information by interactively aggregating the fine-grained time series features of wind direction and wind speed, and then realize the adaptive control of the crane moving speed based on the main component matching fusion features of load and wind force. In this way, it is possible to dynamically adapt to wind changes and reasonably control the moving speed of the crane, thereby improving the working stability and safety of the crane in complex environments.

[0102] Figure 2 FIG. 1 is a block diagram of a crane adaptive control system according to an exemplary embodiment. Figure 2 As shown, the system 200 includes:

[0103] A load value, wind direction and wind speed acquisition module 201 is used to acquire a time queue of load values ​​acquired by a load sensor, and a time queue of wind direction and a time queue of wind speed acquired by an environmental sensor;

[0104] A wireless communication module 202, used to transmit the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed to the adaptive control center through a wireless communication module;

[0105] A time series feature extraction module 203 is used to extract time series features of the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed in the adaptive control center to obtain a load time series associated implicit feature vector, a wind direction time series associated implicit feature vector and a wind speed time series associated implicit feature vector;

[0106] A local fine-grained sequence interaction response aggregation module 204 is used to input the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector into a local fine-grained sequence interaction response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interaction global representation vector;

[0107] The moving speed control instruction determination module 205 is used to determine the moving speed control instruction based on the principal component matching fusion feature between the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector.

[0108] Reference below Figure 3, which shows a schematic diagram of the structure of an electronic device 600 suitable for implementing the embodiment of the present application. The terminal device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0109] like Figure 3 As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0110] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0111] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present application are executed.

[0112] It should be noted that the computer-readable medium mentioned above in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0113] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0114] The computer-readable medium may be included in the electronic device, or may exist independently without being installed in the electronic device.

[0115] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0116] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0117] The modules involved in the embodiments described in this application can be implemented by software or hardware. The name of the module does not limit the module itself in some cases. For example, the test parameter acquisition module can also be described as a "module for acquiring device test parameters corresponding to the target device".

[0118] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0119] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0120] Figure 4 is an application scenario diagram of a crane adaptive control method according to an exemplary embodiment. Figure 4 As shown, in this application scenario, first, a time queue of load values ​​collected by a load sensor is obtained (for example, Figure 4 C1 as shown in the figure), and the time sequence of wind direction collected by the environmental sensor (e.g. Figure 4 C2) and the time sequence of wind speed (e.g. Figure 4 Then, the obtained time queue of load values, time queue of wind direction and time queue of wind speed are input to a server (for example, Figure 4 In S) as shown in , the server is capable of processing the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed based on the crane adaptive control algorithm to determine the control instruction of the moving speed.

[0121] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

[0122] In addition, although each operation is described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or to be performed in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the application. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0123] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely exemplary forms of implementation. With respect to the apparatus in the above-mentioned embodiments, the specific manner in which each module performs the operation has been described in detail in the embodiments related to the method, and will not be described in detail here.

Claims

1. A crane adaptive control method, characterized in that: include: Obtaining a time queue of load values ​​collected by a load sensor, and a time queue of wind direction and a time queue of wind speed collected by an environmental sensor; Transmitting the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed to an adaptive control center through a wireless communication module; In the adaptive control center, time series feature extraction is performed on the time queue of the load value, the time queue of the wind direction, and the time queue of the wind speed to obtain a load time series associated implicit feature vector, a wind direction time series associated implicit feature vector, and a wind speed time series associated implicit feature vector; Inputting the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector into a local fine-grained sequence interaction response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interaction global representation vector; Determine a control instruction for the moving speed based on the principal component matching fusion feature between the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector; The time series feature extraction is performed on the time series of the load value, the time series of the wind direction, and the time series of the wind speed respectively to obtain a load time series associated implicit feature vector, a wind direction time series associated implicit feature vector, and a wind speed time series associated implicit feature vector, including: Inputting the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed into a sequence encoder based on a Bi-LSTM model respectively to obtain the load time series associated implicit feature vector, the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector; Inputting the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector into a local fine-grained sequence interaction response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interaction global representation vector, including: Performing feature decoupling on the implicit feature vector associated with the wind direction time series and the implicit feature vector associated with the wind speed time series to obtain a sequence of wind direction time series local representation vectors and a sequence of wind speed time series local representation vectors; Calculating the local semantic association interaction vectors between each group of corresponding wind direction time series local representation vectors and wind speed time series local representation vectors in the sequence of wind direction time series local representation vectors and the sequence of wind speed time series local representation vectors to obtain a sequence of wind power local time series semantic association interaction vectors; Calculating the fine-grained attention weight of each wind power local temporal semantic association interaction vector in the sequence of wind power local temporal semantic association interaction vectors to obtain a sequence of fine-grained attention weights; Inputting the sequence of fine-grained attention weights into a random dropout module to obtain a sequence of fine-grained attention weights after dropout processing; Taking the sequence of fine-grained attention weights after the deactivation process as the weight distribution, calculating the weighted sum of the sequence of wind power local temporal semantic association interaction vectors to obtain the wind power temporal decoupling reinforcement interaction global representation vector; Calculating the local semantic association interaction vector between each group of corresponding wind direction time series local representation vectors and wind speed time series local representation vectors in the sequence of wind direction time series local representation vectors and the sequence of wind speed time series local representation vectors to obtain a sequence of wind power local time series semantic association interaction vectors, including: Calculate the position point addition, position point subtraction and position point multiplication between the wind direction time series local representation vector and the corresponding wind speed time series local representation vector to obtain a first fusion result, a second fusion result and a third fusion result; After cascading the first fusion result, the second fusion result and the third fusion result into a wind multi-scale fusion feature vector, one-dimensional convolution processing and maximum pooling processing are performed on the vector to obtain the wind local temporal semantic association interaction vector; Calculating the fine-grained attention weight of each wind power local temporal semantic association interaction vector in the sequence of wind power local temporal semantic association interaction vectors to obtain a sequence of fine-grained attention weights, including: The wind local time series semantic association interaction vector is multiplied by a weight coefficient matrix, and then added to the bias vector and passed through a tanh function to obtain a wind local time series semantic association interaction encoding feature vector; Multiplying the wind local temporal semantic association interaction encoding feature vector by the attention score reference vector to obtain the fine-grained attention weight; Determining a control instruction of a moving speed based on the principal component matching fusion feature between the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector includes: Input the load time series associated implicit feature vector and the wind power time series decoupling enhanced interaction global representation vector into a significant fusion network based on feature principal component query matching to obtain a load-wind power principal component collaborative interaction fusion representation vector; Inputting the load-wind principal component collaborative interaction fusion representation vector into a decoder-based mobile speed controller to obtain the control instruction, wherein the control instruction is used to represent a recommended mobile speed decoding value; Inputting the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector into a significant fusion network based on feature principal component query matching to obtain a load-wind principal component collaborative interaction fusion representation vector, including: The load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector are standardized to obtain a standardized load time series associated implicit feature vector and a standardized wind time series decoupling enhanced interaction global representation vector; Respectively calculating the sample covariance matrices of the standardized load time series associated implicit eigenvector and the standardized wind time series decoupled enhanced interaction global representation vector to obtain a load sample covariance matrix and a wind sample covariance matrix; Performing matrix decomposition-based eigenvector extraction on the load sample covariance matrix and the wind sample covariance matrix to obtain a set of load time series associated principal component eigenvectors and a set of wind time series interactive principal component eigenvectors; Inputting the set of the load time series associated principal component feature vectors and the set of the wind time series interactive principal component feature vectors into a maximum approximate query matching network to obtain a set of best matching pairs of the load time series associated principal component feature vectors and the wind time series interactive principal component feature vectors; Input each best matching pair in the set of best matching pairs of the load time series association principal component feature vector and the wind time series interaction principal component feature vector into a fine-grained semantic interaction joint module to obtain a set of load-wind principal component feature vectors; Cascading the set of load-wind principal component feature vectors to obtain the load-wind principal component collaborative interaction fusion representation vector; Among them, each best matching pair in the set of best matching pairs of the load time series association principal component feature vector and the wind time series interaction principal component feature vector is input into a fine-grained semantic interaction joint module to obtain a set of load-wind principal component feature vectors, including: performing multi-dimensional feature fusion on the best matching pairs, and cascading the multi-dimensional feature fusion results into a load-wind principal component multi-scale fusion feature vector, and then performing one-dimensional convolution processing and maximum pooling processing on it to obtain the load-wind principal component feature vector.

2. The crane adaptive control method according to claim 1, characterized in that: Performing matrix decomposition-based feature vector extraction on the load sample covariance matrix and the wind sample covariance matrix to obtain a set of load time series associated principal component feature vectors and a set of wind time series interactive principal component feature vectors, including: Calculate the principal component eigenvalues ​​and principal component direction eigenvectors of the wind sample covariance matrix so that the wind sample covariance matrix multiplied by the principal component direction eigenvector is equal to the principal component eigenvalue multiplied by the principal component direction eigenvector, so as to obtain a set of wind time series interactive principal component eigenvalues ​​and a set of wind time series interactive principal component direction eigenvectors; Using the set of wind time series interactive principal component eigenvalues ​​as diagonal elements to construct a wind time series interactive principal component eigenvalue diagonal matrix; Calculate the product of each wind time series interaction principal component directional eigenvector in the set of the wind time series interaction principal component directional eigenvectors and the diagonal matrix of the wind time series interaction principal component eigenvalues, and then multiply the product by the transposed vector of the wind time series interaction principal component directional eigenvector to obtain the set of the wind time series interaction principal component eigenvectors.

3. The crane adaptive control method according to claim 2, characterized in that: Inputting the set of the load time series associated principal component feature vectors and the set of the wind time series interactive principal component feature vectors into a maximum approximate query matching network to obtain a set of optimal matching pairs of the load time series associated principal component feature vectors and the wind time series interactive principal component feature vectors, including: Calculating the cosine similarity between the i-th load time series associated principal component eigenvector in the set of load time series associated principal component eigenvectors and each wind time series interactive principal component eigenvector in the set of wind time series interactive principal component eigenvectors to obtain a set of cosine similarities; The wind time series interaction principal component eigenvector corresponding to the maximum value in the set of cosine similarities and the i-th load time series association principal component eigenvector form the best matching pair.

4. A crane adaptive control system, using a crane adaptive control method according to any one of claims 1 to 3, characterized in that: include: A load value, wind direction and wind speed acquisition module, used to acquire a time queue of load values ​​collected by a load sensor, and a time queue of wind direction and a time queue of wind speed collected by an environmental sensor; A wireless communication module, used for transmitting the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed to an adaptive control center through the wireless communication module; A time series feature extraction module is used to extract time series features of the time queue of the load value, the time queue of the wind direction and the time queue of the wind speed in the adaptive control center to obtain a load time series associated implicit feature vector, a wind direction time series associated implicit feature vector and a wind speed time series associated implicit feature vector; A local fine-grained sequence interaction response aggregation module is used to input the wind direction time series associated implicit feature vector and the wind speed time series associated implicit feature vector into a local fine-grained sequence interaction response aggregation network based on feature decoupling to obtain a wind power time series decoupling enhanced interaction global representation vector; The control instruction determination module for the moving speed is used to determine the control instruction for the moving speed based on the principal component matching fusion feature between the load time series associated implicit feature vector and the wind time series decoupling enhanced interaction global representation vector.

Images