Intelligent logistics distribution system and method based on Internet of Vehicles system

Abstract

The invention relates to the field of intelligent recommendation, and particularly discloses an intelligent logistics distribution system and method based on an Internet of Vehicles system, which obtains goods label information, position information of a plurality of preset time points in a preset time period and vehicle sensing data information, and deep learning and cloud computing technologies are adopted to carry out semantic analysis and time sequence association on the cargo label information, the position information and the vehicle sensing data information, so that a reasonable vehicle speed value is generated and recommended to a driver. Through the method, the relationship among various data can be better understood by adopting deep learning and cloud computing technologies, so that the recommended vehicle speed value is more accurately generated, intelligent scheduling of logistics distribution is realized, the cargo safety and the vehicle stability are ensured, the distribution route is optimized, and the distribution efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of intelligent control, and more specifically, to an intelligent logistics distribution system and method based on a vehicle networking system. Background Art

[0002] Logistics distribution refers to the process of transporting goods from the production site or warehouse to the consumption site or destination, including transportation, loading and unloading, sorting, packaging, warehousing and other links. The purpose of logistics distribution is to meet customer needs, improve service levels, reduce costs, and increase efficiency. Logistics distribution is an important part of the logistics system and the core content of logistics management.

[0003] During the logistics distribution process, various accidents may occur, affecting the safety and timeliness of goods. For example, due to traffic congestion or accidents, and the driver not paying attention, the vehicle may have faults such as rear-end collisions and scratches, resulting in vehicle damage or fines, and causing damage or delays to the goods. Therefore, during driving, a reasonable vehicle speed is required to ensure the safe transportation of goods.

[0004] In the traditional logistics transportation process, the control of vehicle speed is mainly based on the driver's experience and intuition. However, the driver may be affected by factors such as fatigue, inattention, and personal preferences, resulting in unstable and inconsistent vehicle speeds, unable to achieve the optimal control strategy, thereby increasing the risk of accidents and the decline of transportation efficiency.

[0005] Therefore, an optimized intelligent logistics distribution system based on a vehicle networking system is expected. Summary of the Invention

[0006] To solve the above technical problems, this application is proposed. Embodiments of this application provide an intelligent logistics distribution system and method based on a vehicle networking system, which obtain goods label information, location information at multiple predetermined time points within a predetermined time period, and vehicle sensing data information. Among them, the vehicle sensing data information includes vehicle speed and acceleration, and uses deep learning and cloud computing technologies to perform semantic analysis and temporal correlation on the goods label information, the location information, and the vehicle sensing data information, so as to generate a recommended vehicle speed value for the driver. Through this method, deep learning and cloud computing technologies can be used to better understand the relationships between various data, thereby generating more accurate recommended vehicle speed values, realizing intelligent scheduling of logistics distribution, ensuring the safety of goods and vehicle stability, optimizing the distribution route, and improving the distribution efficiency.

[0007] According to one aspect of this application, an intelligent logistics distribution system based on a vehicle networking system is provided, which includes:

[0008] A cargo vehicle information data acquisition module, which is used to obtain cargo label information, location information at multiple predetermined time points within a predetermined time period, and vehicle sensing data information. Among them, the vehicle sensing data information includes vehicle speed and acceleration;

[0009] A semantic encoding module for cargo label information, which is used to perform word segmentation on the cargo label information and then perform semantic encoding of cargo information to obtain a semantic encoding feature matrix of cargo label information;

[0010] A time dimension arrangement module, which is used to arrange the location information at multiple predetermined time points within the predetermined time period and the vehicle sensing data information into a location information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector according to samples and time dimensions;

[0011] A vehicle information time series correlation module, which is used to perform time series feature extraction and correlation on the location information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector to obtain a vehicle information time series correlation feature matrix;

[0012] A cargo vehicle information fusion module, which is used to fuse the semantic encoding feature matrix of cargo label information and the vehicle information time series correlation feature matrix to obtain a cargo-vehicle information correlation fusion feature matrix, and then perform in-depth correlation feature analysis to obtain a cargo-vehicle information in-depth correlation fusion feature vector;

[0013] A vehicle speed recommendation module, which is used to generate a reasonable vehicle speed value for recommending drivers based on the cargo-vehicle information in-depth correlation fusion feature vector.

[0014] According to another aspect of the present application, an intelligent logistics distribution method based on a vehicle networking system is provided, which includes:

[0015] Obtain cargo label information, location information at multiple predetermined time points within a predetermined time period, and vehicle sensing data information. Among them, the vehicle sensing data information includes vehicle speed and acceleration;

[0016] Perform word segmentation on the cargo label information and then perform semantic encoding of cargo information to obtain a semantic encoding feature matrix of cargo label information;

[0017] Arrange the location information at multiple predetermined time points within the predetermined time period and the vehicle sensing data information into a location information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector according to samples and time dimensions;

[0018] Perform time series feature extraction and correlation on the location information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector to obtain a vehicle information time series correlation feature matrix;

[0019] Fuse the semantic encoding feature matrix of the goods label information and the time-series correlation feature matrix of the vehicle information to obtain a goods-vehicle information association fusion feature matrix, and then perform in-depth association feature analysis to obtain a goods-vehicle information in-depth association fusion feature vector;

[0020] Based on the goods-vehicle information in-depth association fusion feature vector, generate a recommended reasonable vehicle speed value for the driver.

[0021] Compared with the prior art, an intelligent logistics distribution system and method based on a vehicle networking system provided by the present application obtains goods label information, location information at multiple predetermined time points within a predetermined time period, and vehicle sensing data information, where the vehicle sensing data information includes vehicle speed and acceleration, and uses deep learning and cloud computing technologies to perform semantic analysis and time-series correlation on the goods label information, the location information, and the vehicle sensing data information, so as to generate a recommended reasonable vehicle speed value for the driver. Through this method, deep learning and cloud computing technologies can be used to better understand the relationships between various data, so as to more accurately generate the recommended vehicle speed value, realize the intelligent scheduling of logistics distribution, ensure the safety of goods and the stability of vehicles, optimize the distribution route, and improve the distribution efficiency. Description of the Drawings

[0022] By describing the embodiments of the present application in more detail with reference to the drawings, the above and other objects, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application, and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0023] Figure 1 It is a block diagram of an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application.

[0024] Figure 2 It is a schematic structural diagram of an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application.

[0025] Figure 3 It is a block diagram of a semantic encoding module for goods label information in an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application.

[0026] Figure 4 It is a block diagram of a time-series correlation module for vehicle information in an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application.

[0027] Figure 5 It is a block diagram of a goods-vehicle information fusion module in an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application.

[0028] Figure 6 It is a flowchart of an intelligent logistics distribution method based on a vehicle networking system according to an embodiment of the present application. Detailed implementation manners

[0029] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0030] It should be understood that the various steps recorded in the method embodiments of the present disclosure can be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0031] It should be noted that the terms "first", "second", and "third" involved in the embodiments of the present application are only used to distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first", "second", and "third" can be interchanged in a specific order or sequence when permitted. It should be understood that the objects distinguished by "first", "second", and "third" can be interchanged appropriately so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here.

[0032] As shown in the present disclosure and the claims, unless the context clearly indicates an exception, words such as "a", "an", "one", and / or "the" are not specifically singular and may also include plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0033] Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings do not have to be drawn to scale unless otherwise specified.

[0034] Logistics distribution refers to the process of transporting goods from the production place or warehouse to the consumption place or destination, including transportation, loading and unloading, sorting, packaging, warehousing and other links. The purpose of logistics distribution is to meet the needs of customers, improve service levels, reduce costs, and increase efficiency. Logistics distribution is an important part of the logistics system and also the core content of logistics management.

[0035] During the logistics distribution process, various accidents may occur, affecting the safety and timeliness of goods. For example, due to traffic congestion or accidents, and the driver's inattention, the vehicle may experience faults such as rear-end collisions and scratches, resulting in vehicle damage or fines, and causing damage or delays to the goods. Therefore, during the driving process, a reasonable vehicle speed is required to ensure the safe transportation of goods.

[0036] In the traditional logistics transportation process, the control of vehicle speed is mainly based on the driver's experience and intuition. However, the driver may be affected by factors such as fatigue, inattention, and personal preferences, resulting in unstable and inconsistent vehicle speeds, unable to achieve the optimal control strategy, thereby increasing the risk of accidents and the decline of transportation efficiency. Therefore, an optimized intelligent logistics distribution method based on the vehicle networking system is expected, which obtains the goods label information, the position information at multiple predetermined time points within a predetermined time period, and the vehicle sensing data information. Among them, the vehicle sensing data information includes vehicle speed and acceleration, and uses deep learning and cloud computing technologies to perform semantic and temporal correlation analysis on the goods label information, the position information, and the vehicle sensing data information, thereby generating a recommended vehicle speed value for the driver. Through this method, deep learning and cloud computing technologies can be used to better understand the relationships between various data, so as to more accurately generate the recommended vehicle speed value, realize the intelligent scheduling of logistics distribution, ensure the safety of goods and the stability of the vehicle, optimize the distribution route, and improve the distribution efficiency.

[0037] It is worth mentioning that the vehicle networking system refers to the vehicle Internet system, also known as the vehicle network or vehicle interconnection system. It is a system that connects vehicles to the Internet through wireless communication technology, enabling vehicles to achieve real-time communication and data exchange with other vehicles, infrastructure, and cloud servers, etc. The vehicle networking system usually includes components such as in-vehicle communication modules, in-vehicle sensors, in-vehicle computing devices, and cloud servers. Specifically, through the vehicle networking system, vehicles can achieve functions such as vehicle-to-vehicle communication, vehicle-to-infrastructure communication, and vehicle-to-cloud server communication. Such a system can provide various intelligent services and functions for drivers and vehicles, such as real-time traffic information, navigation services, remote monitoring, vehicle diagnosis, vehicle positioning, remote control, etc. Therefore, the development of the vehicle networking system can enhance driving safety, improve traffic efficiency, improve the driving experience, and have a positive impact on fields such as intelligent transportation and intelligent logistics. With the continuous development and popularization of the Internet of Things technology, the vehicle networking system plays an increasingly important role in modern traffic and logistics management.

[0038] Figure 1 It is a block diagram of an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application. Figure 2 It is a schematic diagram of the architecture of an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application. AsFigure 1 and Figure 2 As shown in Figure 2 , the intelligent logistics distribution system 100 based on the vehicle networking system according to an embodiment of the present application includes: a cargo vehicle information data acquisition module 110, configured to obtain cargo label information, multiple predetermined time point position information within a predetermined time period, and vehicle sensing data information, where the vehicle sensing data information includes vehicle speed and acceleration; a cargo label information semantic encoding module 120, configured to perform word segmentation processing on the cargo label information and then perform semantic encoding of the cargo information to obtain a cargo label information semantic encoding feature matrix; a time dimension arrangement module 130, configured to arrange the multiple predetermined time point position information and vehicle sensing data information within the predetermined time period into a position information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector according to samples and time dimensions; a vehicle information time series association module 140, configured to perform time series feature extraction and association on the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector to obtain a vehicle information time series association feature matrix; a cargo vehicle information fusion module 150, configured to fuse the cargo label information semantic encoding feature matrix and the vehicle information time series association feature matrix to obtain a cargo-vehicle information association fusion feature matrix, and then perform deep association feature analysis to obtain a cargo-vehicle information deep association fusion feature vector; and a vehicle speed recommendation module 160, configured to generate a recommended reasonable vehicle speed value for the driver based on the cargo-vehicle information deep association fusion feature vector.

[0039] In an embodiment of the present application, the cargo vehicle information data acquisition module 110 is configured to obtain cargo label information, multiple predetermined time point position information within a predetermined time period, and vehicle sensing data information, where the vehicle sensing data information includes vehicle speed and acceleration. It should be understood that considering that the cargo label information includes attributes of the cargo, such as weight, size, special requirements, etc. The position information can provide the current real-time position of the cargo or the vehicle, helping to monitor and track the transportation process of the cargo. Secondly, it can also reflect the driving route of the cargo or the vehicle, including the starting point, the ending point, and the path information passed through. The vehicle sensing data information reflects the real-time speed of the vehicle, which can be used to monitor the driving condition of the vehicle and the compliance with traffic rules. Specifically, the acceleration provides data on the acceleration and deceleration of the vehicle, which can be used to analyze driving behavior and vehicle performance. Based on this, in the technical solution of the present application, obtaining cargo label information, multiple predetermined time point position information within a predetermined time period, and vehicle sensing data information, where the vehicle sensing data information includes vehicle speed and acceleration, can realize the real-time monitoring and management of the entire logistics distribution process, help improve transportation efficiency, ensure cargo safety, optimize the distribution route, and reduce transportation costs. The analysis and utilization of these data can provide better services and experiences for logistics companies and customers.

[0040] In the embodiment of the present application, the semantic encoding module 120 of the goods label information is configured to perform semantic encoding of the goods information after word segmentation processing of the goods label information to obtain a semantic encoding feature matrix of the goods label information. Figure 3 It is a block diagram of the semantic encoding module of the goods label information in the intelligent logistics distribution system based on the vehicle networking system according to the embodiment of the present application. Specifically, in the embodiment of the present application, as Figure 3 shown, the semantic encoding module 120 of the goods label information includes: a goods information embedding encoding unit 121, configured to perform word segmentation processing on the goods label information and then obtain a plurality of semantic encoding feature vectors of the goods label information through a goods information semantic encoding module including a word embedding layer; and a two-dimensional arrangement unit 122, configured to perform two-dimensional arrangement on the plurality of semantic encoding feature vectors of the goods label information to obtain the semantic encoding feature matrix of the goods label information.

[0041] Specifically, in the embodiment of the present application, the goods information embedding encoding unit 121 is configured to perform word segmentation processing on the goods label information and then obtain a plurality of semantic encoding feature vectors of the goods label information through a goods information semantic encoding module including a word embedding layer. Accordingly, considering that the goods label information consists of multiple words, and there is semantic information and correlation between each word, therefore, in order to better understand each word about the goods information in the goods label information for subsequent better analysis and understanding of these words, in the technical solution of the present application, the goods label information is subjected to word segmentation processing and then passed through a goods information semantic encoding module including a word embedding layer to obtain a plurality of semantic encoding feature vectors of the goods label information. It should be understood that performing word segmentation processing on the goods label information can better understand the meaning and role of each word regarding different goods label information, which is beneficial to capturing the semantic information therein, thereby better expressing the semantic relationship between each word of the goods label information and the associated influence between contexts, so as to provide a richer and more useful feature representation for subsequent data processing and analysis.

[0042] More specifically, in the embodiments of the present application, the goods information embedding and encoding unit includes: a goods label information word segmentation subunit, configured to perform word segmentation on the goods label information to convert the goods label information into a goods label information word sequence composed of multiple words; a goods label information word embedding subunit, configured to use the embedding layer of the goods information semantic encoding module including the word embedding layer to map each word in the goods label information word sequence into a word embedding vector respectively to obtain a sequence of goods label information word embedding vectors; and a goods label information context encoding subunit, configured to use the transformer of the goods information semantic encoding module including the word embedding layer to perform global context semantic encoding based on the transformer idea on the sequence of goods label information word embedding vectors to obtain the plurality of goods label information semantic encoding feature vectors.

[0043] Specifically, in the embodiments of the present application, the two-dimensional arrangement unit 122 is configured to perform two-dimensional arrangement on the plurality of goods label information semantic encoding feature vectors to obtain the goods label information semantic encoding feature matrix. It should be understood that considering that the plurality of goods label information semantic encoding feature vectors contain different information of the goods information semantics, and there is semantic association and mutual influence between the context of each of the goods label information semantic encoding feature vectors. Based on this, in order to comprehensively integrate the semantic features between the goods label information semantic encoding feature vectors, in the technical solution of the present application, the plurality of goods label information semantic encoding feature vectors are arranged two-dimensionally to obtain the goods label information semantic encoding feature matrix. That is, through the two-dimensional arrangement of the plurality of goods label information semantic encoding feature vectors, the semantic features of the plurality of goods label information semantic encoding feature vectors regarding the goods label information can be more comprehensively integrated, so as to obtain a goods label information semantic encoding feature matrix with a richer feature representation.

[0044] In an embodiment of the present application, the time dimension arrangement module 130 is configured to arrange the position information of multiple predetermined time points and the vehicle sensing data information within the predetermined time period into a position information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector according to samples and the time dimension. Correspondingly, considering that the position information and the vehicle sensing data information, including vehicle speed and acceleration, usually have a time series relationship, that is, the position information and the vehicle sensing data information change continuously over time. And different sensing data in the position information and the vehicle sensing data information have different characteristics in terms of time series. Therefore, in order to perform time series feature analysis on each parameter separately to better understand the feature information of each parameter in the time dimension, in the technical solution of the present application, the position information of multiple predetermined time points and the vehicle sensing data information within the predetermined time period are arranged into a position information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector according to samples and the time dimension. It should be understood that arranging the position information of multiple predetermined time points and the vehicle sensing data information within the predetermined time period according to the time dimension helps to extract the features in the time series data of the position information and the vehicle sensing data information. For example, features of different time periods can be extracted by methods such as a sliding window, so as to better capture the dynamic changes and rules of the position information and the vehicle sensing data information, and help the model learn the time series patterns and rules in the data.

[0045] In an embodiment of the present application, the vehicle information time series association module 140 is configured to perform time series feature extraction and association on the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector to obtain a vehicle information time series association feature matrix. Figure 4 It is a block diagram of a vehicle information time series association module in an intelligent logistics distribution system based on a vehicle networking system according to an embodiment of the present application. Specifically, in an embodiment of the present application, as Figure 4 shown, the vehicle information time series association module 140 includes: a vehicle information time series extraction unit 141, configured to obtain a position information time series feature vector, a vehicle speed time series feature vector, and an acceleration time series feature vector by passing the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector through a time series feature extractor based on a one-dimensional convolutional layer; and a time series association unit 142, configured to arrange the position information time series feature vector, the vehicle speed time series feature vector, and the acceleration time series feature vector into the vehicle information time series association feature matrix.

[0046] Specifically, in the embodiment of the present application, the vehicle information time series extraction unit 141 is configured to obtain a position information time series feature vector, a vehicle speed time series feature vector, and an acceleration time series feature vector by passing the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector through a time series feature extractor based on a one-dimensional convolutional layer. It should be understood that considering that the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector all have local time series feature information in the time dimension, and the one-dimensional convolutional layer can effectively capture local features in time series data. Therefore, in order to more accurately understand and analyze the local time series feature information in the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector, in the technical solution of the present application, the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector are passed through a time series feature extractor based on a one-dimensional convolutional layer to obtain a position information time series feature vector, a vehicle speed time series feature vector, and an acceleration time series feature vector. That is, the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector are passed through a time series feature extractor based on a one-dimensional convolutional layer to respectively capture the local time series features of the position information, vehicle speed, and acceleration in the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector, so as to obtain a more representative position information time series feature vector, vehicle speed time series feature vector, and acceleration time series feature vector. In this way, it helps the model better understand the dynamic changes of the data and provides a basis and support for the subsequent recommendation of driving vehicle speed values.

[0047] Specifically, in the embodiment of the present application, the time series correlation unit 142 is configured to arrange the position information time series feature vector, the vehicle speed time series feature vector, and the acceleration time series feature vector into the vehicle information time series correlation feature matrix. Correspondingly, considering that there is a mutual correlation and influence between the position information time series feature vector, the vehicle speed time series feature vector, and the acceleration time series feature vector. Therefore, in order to better understand and utilize the time series relationship between different vehicle information features, in the technical solution of the present application, the position information time series feature vector, the vehicle speed time series feature vector, and the acceleration time series feature vector are arranged into the vehicle information time series correlation feature matrix to integrate the mutual correlation relationship of the position information, the vehicle speed, and the acceleration in the time dimension, so as to obtain a more comprehensive vehicle information time series correlation feature matrix. In this way, the time series correlation between different vehicle information features can be maintained, that is, the corresponding relationship between different vehicle information features in time is retained, so that the model can better understand and utilize the time series relationship between different vehicle information features to recommend reasonable vehicle speed values.

[0048] In the embodiment of the present application, the cargo vehicle information fusion module 150 is configured to fuse the semantic coding feature matrix of the cargo label information and the temporal correlation feature matrix of the vehicle information to obtain a cargo-vehicle information correlation fusion feature matrix, and then perform in-depth correlation feature analysis to obtain a cargo-vehicle information in-depth correlation fusion feature vector. Figure 5 It is a block diagram of the cargo vehicle information fusion module in the intelligent logistics distribution system based on the vehicle networking system according to the embodiment of the present application. Specifically, in the embodiment of the present application, as Figure 5 shown, the cargo vehicle information fusion module 150 includes: a cargo-vehicle information correlation unit 151, configured to fuse the semantic coding feature matrix of the cargo label information and the temporal correlation feature matrix of the vehicle information to obtain the cargo-vehicle information correlation fusion feature matrix; and a deep convolution correlation unit 152, configured to pass the cargo-vehicle information correlation fusion feature matrix through a deep convolution neural network model including multiple convolutional layers to obtain the cargo-vehicle information in-depth correlation fusion feature vector.

[0049] Specifically, in the embodiment of the present application, the cargo-vehicle information correlation unit 151 is configured to fuse the semantic coding feature matrix of the cargo label information and the temporal correlation feature matrix of the vehicle information to obtain the cargo-vehicle information correlation fusion feature matrix. Correspondingly, considering that the semantic coding feature matrix of the cargo label information reflects the information about the cargo in the logistics distribution. The temporal correlation feature matrix of the vehicle information reflects the driving conditions of the vehicle during the logistics distribution process. That is, the cargo label information and the vehicle information are important factors in the logistics distribution process, which means that the cargo information and the vehicle information are often closely related. Based on this, in the technical solution of the present application, the semantic coding feature matrix of the cargo label information and the temporal correlation feature matrix of the vehicle information are fused to obtain the cargo-vehicle information correlation fusion feature matrix. It should be understood that fusing the cargo label information and the vehicle information can integrate information from different sources, better understand the correlation relationship between the two, form a more comprehensive and rich feature representation, and thus help improve the representation ability of the model and the prediction performance of the speed value.

[0050] Specifically, in the embodiments of the present application, the deep convolutional association unit 152 is configured to obtain the cargo-vehicle information deep association fusion feature vector by passing the cargo-vehicle information association fusion feature matrix through a deep convolutional neural network model including multiple convolutional layers. More specifically, in the embodiments of the present application, the deep convolutional neural network model including multiple convolutional layers includes a parallel first convolutional branch structure, a second convolutional branch structure, a third convolutional branch structure, and a fourth convolutional branch structure, and a multi-scale fusion structure connected to the first to fourth convolutional branch structures, wherein the first convolutional branch uses a first convolutional kernel, the second convolutional branch uses a second convolutional kernel, the third convolutional branch uses a third convolutional kernel, and the fourth convolutional branch uses a fourth convolutional kernel, wherein the first convolutional kernel, the second convolutional kernel, the third convolutional kernel, and the fourth convolutional kernel have the same size, and the second convolutional kernel, the third convolutional kernel, and the fourth convolutional kernel have different dilation rates. Accordingly, in order to enhance the feature expression ability of the cargo-vehicle information association fusion feature matrix, further, in the technical solution of the present application, the cargo-vehicle information association fusion feature matrix is passed through a deep convolutional neural network model including multiple convolutional layers to obtain the cargo-vehicle information deep association fusion feature vector. Specifically, the deep convolutional neural network can gradually extract the abstract features in the cargo-vehicle information association fusion feature matrix through multiple convolutional layers, so as to better capture the hierarchical feature representation of the cargo-vehicle information association fusion feature, which helps to improve the model's understanding ability of data. And the deep convolutional neural network has a powerful expression ability and can learn complex non-linear relationships. Through multiple convolutional layers, more complex and abstract feature representations can be gradually constructed, thereby enhancing the model's representation ability and obtaining more abstract and effective feature representations.

[0051] In the embodiments of the present application, the vehicle speed recommendation module 160 is configured to generate a reasonable vehicle speed value recommended for the driver based on the cargo-vehicle information deep association fusion feature vector. Specifically, in the embodiments of the present application, the vehicle speed recommendation module includes: a feature optimization unit configured to perform multi-scale displacement entropy coupling regression optimization on the cargo-vehicle information deep association fusion feature vector to obtain an optimized cargo-vehicle information deep association fusion feature vector; and a decoding unit configured to obtain a decoded value by passing the optimized cargo-vehicle information deep association fusion feature vector through a decoder, and the decoded value is used to represent a reasonable vehicle speed value recommended for the driver.

[0052] Specifically, the feature optimization unit is used to perform multi-scale displacement entropy coupling regression optimization on the deep association fusion feature vector of the goods-vehicle information to obtain an optimized deep association fusion feature vector of the goods-vehicle information. In particular, in the above technical solution, on the one hand, the deep association fusion feature vector of the goods-vehicle information is a feature extracted based on the goods label information, the position information at a predetermined time point, and the vehicle sensing data information. After feature extraction and processing, a high-dimensional feature matrix is formed. The high-dimensional deep association fusion feature vector of the goods-vehicle information can more comprehensively describe the association features between the goods and vehicle information, improve the feature representation ability, and help to more accurately recommend reasonable vehicle speed values. However, at the same time, the high-dimensional deep association fusion feature vector of the goods-vehicle information will increase the computational complexity, require more computational resources and time to process and train the model, and may lead to a decline in system performance. Excessively high dimensions are also prone to overfitting problems, that is, the model performs well on the training set but poorly on the test set, affecting the generalization ability of the model. Based on this, in the technical solution of this application, multi-scale displacement entropy coupling regression optimization is performed on the deep association fusion feature vector of the goods-vehicle information to obtain an optimized deep association fusion feature vector of the goods-vehicle information.

[0053] Specifically, in the embodiment of this application, the feature optimization unit is used to: calculate the autoregressive covariance matrix of the deep association fusion feature vector of the goods-vehicle information, and perform key factor analysis on the autoregressive covariance matrix of the deep association fusion feature vector of the goods-vehicle information to obtain a set of deep association fusion feature basic feature coding vectors of the goods-vehicle information, which is expressed by the formula:

[0054]

[0055]

[0056] where V represents the deep association fusion feature vector of the goods-vehicle information, T represents the transpose of the matrix, M z represents the autoregressive covariance matrix, U represents a set of deep association fusion feature basic feature coding vectors of the goods-vehicle information, v1, v2, v m respectively represent the first, second, and mth deep association fusion feature basic feature coding vectors of the goods-vehicle information, Λ represents the deep association fusion feature diagonal matrix after key factor analysis, and λ1, λ m respectively represent the first and mth eigenvalues on the diagonal of the deep association fusion feature diagonal matrix of the goods-vehicle information.

[0057] That is, as a classic linear dimensionality reduction technique, principal component analysis (PCA) can decouple the linear correlation between features by analyzing the covariance structure of the autoregressive covariance matrix of the feature vectors obtained from the in-depth association and fusion of cargo-vehicle information, and extract the principal component direction with the largest variance in the data, thereby significantly reducing the feature dimension while retaining the core information. Specifically, through PCA, the high-dimensional fusion features are mapped to a low-dimensional orthogonal space, maximizing the variance after data projection, filtering out noise interference and secondary features, and retaining the key temporal correlation patterns between the cargo transportation state and vehicle dynamics parameters (such as the strong correlation between rapid acceleration and the risk of cargo jolting), thus providing a more discriminative and compact feature representation for the subsequent vehicle speed recommendation module.

[0058] Specifically, in the embodiment of the present application, the feature optimization unit is further configured to: input the set of basic feature encoding vectors of the in-depth association and fusion features of cargo-vehicle information into a sequence encoder based on a forward LSTM model to obtain a set of context-associated encoding vectors of the basic features of the in-depth association and fusion features of cargo-vehicle information, which is represented by the formula:

[0059] F = LSTM([v1, v2, …, v m ) = [s1, s2, …, s m

[0060] where LSTM represents the forward LSTM model, F represents the set of context-associated encoding vectors of the basic features of the in-depth association and fusion features of cargo-vehicle information, and s1, s2, s m represent the first, second, and m-th context-associated encoding vectors of the basic features of the in-depth association and fusion features of cargo-vehicle information.

[0061] That is, due to the sequence modeling mechanism of the forward LSTM, which can analyze the information organizational structure implicit in the sorting of basic features. The globally ranked features may determine the macroscopic framework of the transportation strategy, while the locally ranked features reflect the detailed fluctuations of the dynamic road conditions. Therefore, through the non-linear context dependence modeling of the basic feature encoding vectors of the in-depth association and fusion features of cargo-vehicle information by the LSTM model in the present application, the progressive interaction logic between the cargo sensitivity and the vehicle operating state in the spatio-temporal dimension can be captured, and the discrete basic features of the in-depth association and fusion features of cargo-vehicle information can be transformed into a set of context-associated encoding vectors of the basic features of the in-depth association and fusion features of cargo-vehicle information with temporal perception ability, thereby revealing the causal chain hidden in the high-dimensional features. On the one hand, LSTM filters out redundant temporal noise through the gating mechanism and strengthens key dynamic features. On the other hand, the context-associated encoding vectors of the basic features of the in-depth association and fusion features of cargo-vehicle information generated by sequence encoding can provide a decision basis for subsequent processing.

[0062] ​Specifically, in the embodiments of the present application, the feature optimization unit is further configured to: calculate the bit-by-bit dislocation entropy between each corresponding cargo-vehicle information depth association fusion feature basic feature context association coding vector and cargo-vehicle information depth association fusion feature basic feature coding vector in the set of cargo-vehicle information depth association fusion feature basic feature context association coding vectors and the set of cargo-vehicle information depth association fusion feature basic feature coding vectors to obtain a set of bit-by-bit dislocation entropy, which is represented by the formula:

[0063]

[0064] wherein, ∧ represents a logical operator, w represents a bit-by-bit dislocation comparison function, v i represents the i-th cargo-vehicle information depth association fusion feature basic feature coding vector, represents the feature value at the j-th position of the i-th cargo-vehicle information depth association fusion feature basic feature coding vector, s i represents the i-th cargo-vehicle information depth association fusion feature basic feature context association coding vector, represents the feature value at the j-th position of the i-th cargo-vehicle information depth association fusion feature basic feature context association coding vector, ε represents a predetermined threshold, r i represents the i-th bit-by-bit dislocation comparison feature vector, represents the feature value at the j-th position of the i-th bit-by-bit dislocation comparison feature vector, L represents the length of the i-th bit-by-bit dislocation comparison feature vector, e i represents the i-th bit-by-bit dislocation entropy.

[0065] That is, by quantifying the information perturbation pattern at the binary bit level, the bit-by-bit dislocation entropy can reveal the semantic-level association reconstruction that occurs in the context modeling process of the feature basic features, especially for the complex non-linear coupling relationship between the cargo attributes and vehicle dynamics parameters in the logistics scenario. In this way, by constructing an information theory-oriented optimization signal to capture the change intensity, direction, and probability distribution of the feature binary bits, the information gain or loss of the context association modeling for the original feature space can be measured, thereby providing a sensitive and fine-grained differentiation basis for subsequent weight adjustment. On the one hand, the bit-by-bit dislocation entropy enhances the model's perception ability of the cargo sensitivity threshold by quantifying the dynamic differences at the information level, avoiding the optimization deviation caused by traditional metrics ignoring the binary bit semantics; on the other hand, the entropy value set guides the weight allocation strategy to preferentially strengthen the high information entropy region by mapping the complexity gradient of the feature changes.

[0066] Specifically, in the embodiments of the present application, the feature optimization unit is further configured to: perform weight processing on the set of bit-by-bit dislocation entropy to obtain a set of bit-by-bit repositioning entropy adjustment factors, which is represented by the formula:

[0067] a i = softmax(e i )

[0068] where softmax represents the normalized exponential function, and a i represents the i-th bitwise repositioning entropy adjustment factor.

[0069] That is, the present application introduces the Softmax function, which transforms the entropy value difference into a probability distribution form through exponential mapping, enabling the basic features corresponding to high entropy values to obtain significantly higher optimization weights. At the same time, the smoothness constraint avoids the emergence of extreme weights. Its execution purpose is to construct a dynamic adaptive feature importance evaluation mechanism. By amplifying the weight distribution in the high entropy value region, it enhances the response sensitivity of the model to key information perturbations, while suppressing the noise interference in the low entropy value region, thereby providing a decision-making basis that conforms to the causal logic of the logistics scenario for subsequent optimization.

[0070] Specifically, in the embodiment of the present application, the feature optimization unit is further configured to: based on the set of bitwise repositioning entropy adjustment factors, fuse the set of basic feature encoding vectors of the cargo-vehicle information depth association fusion feature to obtain an optimized cargo-vehicle information depth association fusion feature vector, which is represented by the formula:

[0071]

[0072] where v f represents the optimized cargo-vehicle information depth association fusion feature vector.

[0073] That is, through the differential adjustment of the bitwise repositioning entropy adjustment factor for the basic features of the cargo-vehicle information depth association fusion feature, it strengthens the marginal contribution of the high information entropy region to the vehicle speed recommendation decision, while weakening the interference of redundant features, so that the fused optimized cargo-vehicle information depth association fusion feature vector can more accurately map the non-linear constraint relationship between the physical characteristics of the cargo and the dynamic parameters of the vehicle, enhancing the robustness of the model to complex logistics scenarios.

[0074] Specifically, the decoding unit is configured to obtain a decoded value by decoding the optimized cargo-vehicle information depth association fusion feature vector through a decoder. The decoded value is used to represent a reasonable vehicle speed value recommended for the driver. That is, by performing depth convolution on the cargo-vehicle information depth association fusion feature vector obtained, decoding regression is performed to generate a reasonable vehicle speed value recommended for the driver based on the comprehensive representation of the cargo and vehicle information. Through this method, the relationship between various data can be better understood, so as to more accurately generate the recommended vehicle speed value, realize the intelligent scheduling of logistics distribution, thereby ensuring the safety of goods and the stability of the vehicle, optimizing the distribution route and improving the distribution efficiency.

[0075] Specifically, in the embodiments of the present application, the vehicle speed recommendation module is configured to: use the decoder to perform decoding regression on the deeply associated fusion feature vector of the cargo-vehicle information to obtain the decoded value; wherein, the decoding formula is: where X is the deeply associated fusion feature vector of the cargo-vehicle information, Y is the decoded value, and W is the weight matrix, represents matrix multiplication.

[0076] In summary, the intelligent logistics distribution system 100 based on the vehicle networking system according to the embodiments of the present application is elucidated. It acquires cargo label information, location information at multiple predetermined time points within a predetermined time period, and vehicle sensing data information. Among them, the vehicle sensing data information includes vehicle speed and acceleration, and uses deep learning and cloud computing technologies to perform semantic analysis and temporal correlation on the cargo label information, the location information, and the vehicle sensing data information, so as to generate a recommended reasonable vehicle speed value for the driver. Through this method, deep learning and cloud computing technologies can be used to better understand the relationships between various data, so as to more accurately generate the recommended vehicle speed value, realize the intelligent scheduling of logistics distribution, ensure the safety of goods and the stability of the vehicle, optimize the distribution route, and improve the distribution efficiency.

[0077] Figure 6 is a flowchart of the intelligent logistics distribution method based on the vehicle networking system according to the embodiments of the present application. As Figure 6 shown, the intelligent logistics distribution method based on the vehicle networking system according to the embodiments of the present application includes: S110, acquiring cargo label information, location information at multiple predetermined time points within a predetermined time period, and vehicle sensing data information, where the vehicle sensing data information includes vehicle speed and acceleration; S120, performing word segmentation processing on the cargo label information and then performing semantic encoding of the cargo information to obtain a cargo label information semantic encoding feature matrix; S130, arranging the location information at multiple predetermined time points and the vehicle sensing data information within the predetermined time period in the sample and time dimensions to form a location information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector; S140, performing time series feature extraction and correlation on the location information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector to obtain a vehicle information time series correlation feature matrix; S150, fusing the cargo label information semantic encoding feature matrix and the vehicle information time series correlation feature matrix to obtain a cargo-vehicle information associated fusion feature matrix, and then performing deep association feature analysis to obtain a deeply associated fusion feature vector of the cargo-vehicle information; and S160, generating a recommended reasonable vehicle speed value for the driver based on the deeply associated fusion feature vector of the cargo-vehicle information.

[0078] Here, those skilled in the art can understand that the specific operations of each step in the above intelligent logistics distribution method based on the vehicle networking system have been described in detail in the description of the intelligent logistics distribution system based on the vehicle networking system above with reference to Figures 1 to 5 and thus, the repeated description thereof will be omitted.

[0079] The embodiments of the present disclosure have been described above. The above description is exemplary and not exhaustive, and is also not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other ordinary skilled in the art to understand the embodiments disclosed herein.

Claims

1. An intelligent logistics distribution system based on a vehicle networking system, characterized in that, Including: A cargo vehicle information data acquisition module, configured to obtain cargo label information, location information at multiple predetermined time points within a predetermined time period, and vehicle sensing data information, wherein the vehicle sensing data information includes vehicle speed and acceleration; A cargo label information semantic encoding module, configured to perform word segmentation on the cargo label information and then perform semantic encoding of the cargo information to obtain a cargo label information semantic encoding feature matrix; A time dimension arrangement module, configured to arrange the location information at multiple predetermined time points within the predetermined time period and the vehicle sensing data information into a location information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector according to samples and time dimensions; A vehicle information time series association module, configured to perform time series feature extraction and association on the location information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector to obtain a vehicle information time series association feature matrix; A cargo vehicle information fusion module, configured to fuse the cargo label information semantic encoding feature matrix and the vehicle information time series association feature matrix to obtain a cargo-vehicle information association fusion feature matrix, and then perform deep association feature analysis to obtain a cargo-vehicle information deep association fusion feature vector; A vehicle speed recommendation module, configured to generate a recommended reasonable vehicle speed value for the driver based on the cargo-vehicle information deep association fusion feature vector.

2. The intelligent logistics distribution system based on the vehicle networking system according to claim 1, characterized in that, The cargo label information semantic encoding module includes: A cargo information embedding encoding unit, configured to perform word segmentation on the cargo label information and then obtain a plurality of cargo label information semantic encoding feature vectors through a cargo information semantic encoding module including a word embedding layer; A two-dimensional arrangement unit, configured to perform two-dimensional arrangement on the plurality of cargo label information semantic encoding feature vectors to obtain the cargo label information semantic encoding feature matrix.

3. The intelligent logistics distribution system based on the vehicle networking system according to claim 2, characterized in that, The cargo information embedding encoding unit includes: A cargo label information word segmentation subunit, configured to perform word segmentation on the cargo label information to convert the cargo label information into a cargo label information word sequence composed of multiple words; A cargo label information word embedding subunit, configured to use the embedding layer of the cargo information semantic encoding module including the word embedding layer to map each word in the cargo label information word sequence into a word embedding vector respectively to obtain a sequence of cargo label information word embedding vectors; A cargo label information context encoding subunit, configured to perform global context semantic encoding based on the idea of a transformer on the sequence of cargo label information word embedding vectors through the transformer of the cargo information semantic encoding module including the word embedding layer to obtain the plurality of cargo label information semantic encoding feature vectors.

4. The intelligent logistics distribution system based on the vehicle networking system according to claim 3, wherein The vehicle information time series association module includes: A vehicle information time series extraction unit, configured to obtain a location information time series feature vector, a vehicle speed time series feature vector, and an acceleration time series feature vector by passing the location information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector through a time series feature extractor based on a one-dimensional convolutional layer; A temporal correlation unit, configured to arrange the position information temporal feature vector, the vehicle speed temporal feature vector, and the acceleration temporal feature vector into the vehicle information temporal correlation feature matrix.

5. The intelligent logistics distribution system based on the vehicle networking system according to claim 4, wherein The cargo vehicle information fusion module includes: A cargo-vehicle information association unit, configured to fuse the cargo label information semantic coding feature matrix and the vehicle information temporal correlation feature matrix to obtain the cargo-vehicle information association fusion feature matrix; A deep convolution association unit, configured to pass the cargo-vehicle information association fusion feature matrix through a deep convolution neural network model including multiple convolutional layers to obtain the cargo-vehicle information deep association fusion feature vector.

6. The intelligent logistics distribution system based on the vehicle networking system according to claim 5, characterized in that The deep convolution neural network model including multiple convolutional layers includes parallel first, second, third, and fourth convolutional branch structures, and a multi-scale fusion structure connected to the first to fourth convolutional branch structures. Among them, the first convolutional branch uses a first convolutional kernel, the second convolutional branch uses a second convolutional kernel, the third convolutional branch uses a third convolutional kernel, and the fourth convolutional branch uses a fourth convolutional kernel. Among them, the first convolutional kernel, the second convolutional kernel, the third convolutional kernel, and the fourth convolutional kernel have the same size, and the second convolutional kernel, the third convolutional kernel, and the fourth convolutional kernel have different dilation rates.

7. The intelligent logistics distribution system based on the vehicle networking system according to claim 6, wherein, The vehicle speed recommendation module includes: A feature optimization unit, configured to perform multi-scale displacement entropy coupling regression optimization on the cargo-vehicle information deep association fusion feature vector to obtain an optimized cargo-vehicle information deep association fusion feature vector; A decoding unit, configured to obtain a decoded value by passing the optimized cargo-vehicle information deep association fusion feature vector through a decoder, and the decoded value is used to represent a reasonable vehicle speed value recommended to the driver.

8. The intelligent logistics distribution system based on the vehicle networking system according to claim 7, wherein, The feature optimization unit is configured to: Calculate the autoregressive covariance matrix of the cargo-vehicle information deep association fusion feature vector, and perform key factor analysis on the autoregressive covariance matrix of the cargo-vehicle information deep association fusion feature vector to obtain a set of cargo-vehicle information deep association fusion feature basic feature coding vectors; Input the set of cargo-vehicle information deep association fusion feature basic feature coding vectors into a sequence encoder based on a forward LSTM model to obtain a set of cargo-vehicle information deep association fusion feature basic feature context association coding vectors; Calculate the bit-by-bit misalignment entropy between each corresponding cargo-vehicle information deep association fusion feature basic feature context association coding vector and cargo-vehicle information deep association fusion feature basic feature coding vector in the set of cargo-vehicle information deep association fusion feature basic feature context association coding vectors and the set of cargo-vehicle information deep association fusion feature basic feature coding vectors to obtain a set of bit-by-bit misalignment entropy; Perform a weighting process on the set of bit-by-bit misalignment entropy to obtain a set of bit-by-bit repositioning entropy adjustment factors; Based on the set of bit-by-bit relocation entropy adjustment factors, fuse the set of basic feature encoding vectors of the deep association fusion features of the cargo-vehicle information to obtain an optimized deep association fusion feature vector of the cargo-vehicle information.

9. An intelligent logistics distribution method based on a vehicle networking system, characterized in that, Including: Obtain the cargo label information, the position information at multiple predetermined time points within a predetermined time period, and the vehicle sensing data information, where the vehicle sensing data information includes vehicle speed and acceleration; After performing word segmentation processing on the cargo label information, perform semantic encoding of the cargo information to obtain a semantic encoding feature matrix of the cargo label information; Arrange the position information at multiple predetermined time points within the predetermined time period and the vehicle sensing data information in accordance with the sample and time dimensions into a position information time series input vector, a vehicle speed time series input vector, and an acceleration time series input vector; Perform time series feature extraction and association on the position information time series input vector, the vehicle speed time series input vector, and the acceleration time series input vector to obtain a vehicle information time series association feature matrix; Fuse the semantic encoding feature matrix of the cargo label information and the vehicle information time series association feature matrix to obtain a cargo-vehicle information association fusion feature matrix, and then perform deep association feature analysis to obtain a deep association fusion feature vector of the cargo-vehicle information; Based on the deep association fusion feature vector of the cargo-vehicle information, generate a recommended reasonable vehicle speed value for the driver.

10. The intelligent logistics distribution method based on the vehicle networking system according to claim 9, characterized in that After performing word segmentation processing on the cargo label information, perform semantic encoding of the cargo information to obtain a semantic encoding feature matrix of the cargo label information, including: After performing word segmentation processing on the cargo label information, pass it through a cargo information semantic encoding module including a word embedding layer to obtain multiple semantic encoding feature vectors of the cargo label information; Arrange the multiple semantic encoding feature vectors of the cargo label information in two dimensions to obtain the semantic encoding feature matrix of the cargo label information.

Images