Data transmission method and device, equipment and storage medium

Abstract

The invention discloses a data transmission method and device, equipment and a storage medium, and relates to the technical field of data transmission. The method comprises the following steps: preprocessing collected initial data to obtain data to be transmitted, and monitoring network quality in a data transmission process in real time to obtain a real-time network condition; compressing the to-be-transmitted data according to the data characteristics of the to-be-transmitted data and the real-time network condition to obtain candidate data; performing feature extraction on the candidate data according to data features to obtain target feature data; and determining a data transmission strategy according to the network topology structure, the data characteristics and the real-time network condition of the edge device, and transmitting the target characteristic data to the target device based on the data transmission strategy. According to the technical scheme, data compression, feature extraction and transmission strategy determination are dynamically carried out according to the real-time network condition, the data transmission problem caused by the unstable network environment can be reduced, and the data transmission efficiency and reliability are effectively improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, in particular to the field of data transmission technology, and specifically to a data transmission method, apparatus, device and storage medium. Background Art

[0002] In recent years, the field of Internet of Things and edge computing has shown a rapid development trend. With the widespread deployment of various smart devices and sensors, the number of devices connected to the Internet of Things has grown exponentially, and massive amounts of data are constantly generated in various scenarios, such as equipment operation data in industrial production, vehicle and road condition information in smart transportation, and various environmental and device status data in smart home environments. Edge computing, as a technical model that moves computing and data storage to the edge of the network, has emerged and attracted much attention. It aims to reduce the delay in data transmission to the cloud, save network bandwidth, and enhance data privacy protection.

[0003] In this development context, edge devices play an increasingly critical role, as they are responsible for collecting, processing, and transmitting large amounts of data. However, edge devices face many challenges in data processing and transmission. Due to the instability of the network environment, problems such as network delays, packet loss, and bandwidth limitations often occur, which affects the efficiency and reliability of data transmission. For example, in intelligent traffic monitoring scenarios, if the network delay is too high, the monitoring data may not be transmitted to the backend for analysis and processing in a timely manner, which in turn affects the real-time control and response decisions of traffic conditions. Summary of the invention

[0004] The present application provides a data transmission method, apparatus, device and storage medium to improve the efficiency and reliability of data transmission.

[0005] According to one aspect of the present application, a data transmission method is provided, the method comprising:

[0006] Pre-process the collected initial data to obtain the data to be transmitted, and monitor the network quality of the data transmission process in real time to obtain the real-time network status;

[0007] According to the data characteristics of the data to be transmitted and the real-time network status, compressing the data to be transmitted to obtain candidate data;

[0008] According to the data characteristics, feature extraction is performed on the candidate data to obtain target feature data;

[0009] A data transmission strategy is determined according to the network topology of the edge device, the data characteristics and the real-time network status, and the target feature data is transmitted to the target device based on the data transmission strategy.

[0010] According to another aspect of the present application, a data transmission device is provided, the device comprising:

[0011] The data acquisition module is used to pre-process the collected initial data to obtain the data to be transmitted, and to monitor the network quality of the data transmission process in real time to obtain the real-time network status;

[0012] A data compression module, used for compressing the data to be transmitted according to the data characteristics of the data to be transmitted and the real-time network status to obtain candidate data;

[0013] A feature extraction module, used to extract features from the candidate data according to the data characteristics to obtain target feature data;

[0014] The strategy determination module is used to determine the data transmission strategy according to the network topology of the edge device, the data characteristics and the real-time network status, and transmit the target feature data to the target device based on the data transmission strategy.

[0015] According to another aspect of the present application, an electronic device is provided, the electronic device comprising:

[0016] one or more processors;

[0017] A memory for storing one or more programs;

[0018] When the one or more programs are executed by the one or more processors, the one or more processors implement any one of the data transmission methods provided in the embodiments of the present application.

[0019] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, any one of the data transmission methods provided in the embodiments of the present application is implemented.

[0020] According to another aspect of the present application, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the computer program implements any one of the data transmission methods provided in the embodiments of the present application.

[0021] This application obtains the data to be transmitted by preprocessing the collected initial data, and monitors the network quality of the data transmission process in real time to obtain the real-time network status; compresses the data to be transmitted according to the data characteristics of the data to be transmitted and the real-time network status to obtain candidate data; extracts features from the candidate data according to the data characteristics to obtain target feature data; determines the data transmission strategy according to the network topology, data characteristics and real-time network status of the edge device, and transmits the target feature data to the target device based on the data transmission strategy. The above technical solution can reduce data transmission problems caused by unstable network environment and effectively improve the efficiency and reliability of data transmission by dynamically compressing data, extracting features and determining transmission strategies according to real-time network conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a flow chart of a data transmission method provided according to Embodiment 1 of the present application;

[0023] Figure 2 is a flow chart of a data transmission method provided according to Embodiment 2 of the present application;

[0024] Figure 3 is a structural schematic diagram of a data transmission device provided according to Embodiment 3 of the present application;

[0025] Figure 4 It is a structural diagram of an electronic device that implements the data transmission method of an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] In addition, it should be noted that in the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of relevant data such as initial data and data to be transmitted are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0029] Embodiment 1

[0030] Figure 1 This is a flowchart of a data transmission method provided according to the first embodiment of the present application. This embodiment is applicable to edge devices performing network data transmission and can be executed by a data transmission device. The data transmission device can be implemented in the form of hardware and / or software. The data transmission device can be configured in a computer device, such as an edge device. Figure 1 As shown, the method includes:

[0031] S110, pre-processing the collected initial data to obtain data to be transmitted, and monitoring the network quality of the data transmission process in real time to obtain real-time network status.

[0032] In this embodiment, initial data refers to the raw data collected during the data collection phase, which is usually data obtained from device sensors, user input, or other data sources without any processing or modification. Preprocessing refers to a series of operations performed on the initial data, aiming to convert the data into a format suitable for further analysis, processing and transmission; it may include steps such as denoising, formatting, standardization, and removal of redundant information, with the aim of improving the quality of the data and reducing the burden during transmission. Data to be transmitted refers to data that has been preprocessed and is ready to enter the data transmission process; it is the final data set extracted and prepared from the initial data. Real-time network status refers to the current state or performance of the network during data transmission, which may include factors such as bandwidth, latency, packet loss rate, signal strength, and network congestion.

[0033] Optionally, the collection of initial data of edge devices is mainly achieved by equipped various sensors, cameras and other equipment. For example, in the intelligent traffic monitoring scenario, the built-in camera can collect image information of vehicles on the road, and the sensor can obtain relevant data such as vehicle speed, vehicle count, accident detection, etc.; in the smart home scenario, the temperature and humidity sensor can collect the temperature and humidity data of the environment, and the door and window sensor can obtain information such as the door and window switch status.

[0034] Furthermore, the collected raw data often has inconsistent formats and contains noise, so preliminary preprocessing is required, which includes operations such as format conversion and data cleaning. Format conversion aims to convert the data collected by different sensors into a format that is convenient for subsequent processing and transmission, such as unifying image data into a specific image format standard. Data cleaning is to remove data that is obviously wrong, repeated, or inconsistent with actual logic, ensure the accuracy and integrity of the data, lay a good foundation for subsequent data processing and transmission, and avoid adverse effects on the entire data transmission process due to raw data quality issues.

[0035] Optional, real-time network quality monitoring is a key link to ensure effective data transmission. The real-time status of the network can be mastered by monitoring the following series of indicators: Delay indicators include round-trip time (RTT), which is the time required for data to travel from the sender to the receiver and then back to the sender, and one-way delay (OWD), which is the time required for data to travel from the sender to the receiver. High latency may affect the timeliness of data transmission. For example, in telemedicine scenarios, if the monitoring data transmission delay is too high, it will delay the diagnosis of the disease; packet loss rate refers to the proportion of data packets that fail to successfully reach the receiver within a certain period of time to the total number of data packets sent. A high packet loss rate will cause data loss, affecting data integrity and subsequent analysis and processing; bandwidth is the available network transmission capacity, usually measured in bits per second. Insufficient bandwidth will limit the speed and amount of data transmission. For example, when transmitting large amounts of high-definition video data, low bandwidth may cause freezes; signal strength is It is particularly important in wireless networks, usually measured in decibel milliwatts. Weak signal strength may lead to unstable connections and data transmission errors. In addition, there is jitter, which is the fluctuation in the arrival time of data packets, which may cause data packets to arrive out of the expected order, affecting the orderly processing of data. Network availability reflects the proportion of time that network services can be used normally. Network congestion reflects the phenomenon that performance is degraded due to excessive data traffic in the network. The error rate represents the proportion of errors that occur during data transmission. Connection stability is related to the stability of the network connection, including the frequency and duration of connection interruptions. Network utilization reflects the use of network resources, such as link utilization, interface utilization, etc.

[0036] It is understandable that edge devices can use the results of the above monitoring indicators to comprehensively evaluate the impact of the current network environment on data transmission, and then provide a strong basis for the dynamic adjustment of subsequent data transmission strategies, ensuring that data transmission can adapt to changes in the network environment and improve the stability and reliability of transmission.

[0037] S120 . According to the data characteristics of the data to be transmitted and the real-time network status, compress the data to be transmitted to obtain candidate data.

[0038] In this embodiment, data characteristics refer to the inherent characteristics of the data to be transmitted, such as the size, format, type, importance, redundancy, change rate, etc. of the data; different data characteristics require different processing, compression and transmission methods. Compression processing refers to the process of reducing the size of data through a certain algorithm; compression can be lossless (maintaining the integrity and accuracy of the data) or approximate compression (sacrificing some data accuracy to further reduce the data volume). Candidate data refers to data obtained after compression processing; these data have been preliminarily optimized and can be further adjusted or processed according to network conditions or transmission strategies so as to select the most suitable method for transmission.

[0039] Exemplarily, a lossless compression algorithm or an approximate compression algorithm may be used to compress the data to be transmitted to obtain candidate data.

[0040] In this embodiment, the lossless compression algorithm refers to a method that does not lose any information during the compression process; the data compressed by this algorithm can be completely restored to the state of the original data when decompressed; that is, the data before and after compression are completely consistent, without any loss of information; the lossless compression algorithm may include Huffman tree coding, LZ77 (Lempel-Ziv 1977) algorithm, etc. Approximate compression algorithm refers to a method of compressing data by discarding part of the information, and the compressed data cannot be completely restored to the original data state, and usually there may be a certain amount of information loss; the lost information is usually perceptually unimportant or redundant data, thereby reducing the file size; the approximate compression algorithm may include vector quantization, wavelet transform, etc.

[0041] S130: Extract features from the candidate data according to data characteristics to obtain target feature data.

[0042] In this embodiment, feature extraction refers to the process of extracting representative, simplified features that can express key information of the data from candidate data; feature extraction can be achieved by using various machine learning, deep learning methods, or based on predefined rules; the goal is to retain the most important features of the data for more efficient transmission and further processing. Target feature data refers to the final data obtained after feature extraction for data transmission; these data contain the key information and features of the original data, can effectively represent the meaning of the original data, and are usually more compact and easier to transmit than the original data.

[0043] It should be noted that feature extraction can be feature extraction based on domain knowledge. For example, for image data, features such as color, texture, and shape can be extracted. In medical image analysis, the shape and color features of the lesion area can be extracted to assist doctors in judging the condition. For time series data, statistical features such as mean, variance, and autocorrelation coefficient can be extracted. In industrial automation scenarios, these statistical features of machine operation data can be analyzed to determine whether the equipment is in an abnormal state.

[0044] It is understandable that this feature extraction method based on domain knowledge can fully combine the actual application needs, extract the most valuable features, and help improve the accuracy of subsequent data analysis and decision-making. However, it requires in-depth understanding and professional knowledge reserves of the corresponding fields, and the feature extraction methods in different fields are less universal and need to be specially designed and adjusted for specific scenarios.

[0045] Optionally, the data characteristics include data types; accordingly, based on the data types in the data characteristics, a target feature extraction method is determined from candidate feature extraction methods; the candidate feature extraction methods include machine learning algorithms and deep learning algorithms; deep learning algorithms include convolutional neural networks and recurrent neural networks; the target feature extraction method is used to extract features from the candidate data to obtain target feature data.

[0046] In this embodiment, data type refers to the classification of data, which determines the storage and operation mode of data; data types may include numerical data, categorical data, time series data, image data, text data, etc. Candidate feature extraction methods refer to a set of potential methods that can be used to extract features from raw data; these methods generally include various machine learning algorithms and deep learning algorithms, and appropriate feature extraction methods can be selected according to data characteristics (such as data types). Machine learning algorithms refer to algorithms that infer models and make predictions by learning existing data (training data). Machine learning algorithms may include decision trees, support vector machines, k-nearest neighbors, random forests, and other algorithms. Deep learning algorithms refer to a branch of machine learning that uses multi-layer neural networks to automatically learn the feature representation of data; deep learning algorithms are particularly suitable for processing complex data types, such as images, audio, and text data. Convolutional neural networks refer to a special type of deep learning model, mainly used for image data processing. Recurrent neural networks refer to a type of deep learning model suitable for sequence data (such as time series, text). The target feature extraction method refers to the data feature extraction technology that is most suitable for a specific task selected from the candidate feature extraction methods; when selecting a feature extraction method, it is necessary to decide based on the type of data, task requirements (such as classification, regression, etc.) and performance requirements of the algorithm.

[0047] For example, when using a convolutional neural network (CNN) to extract features from image data in intelligent traffic monitoring, first build a CNN model containing structures such as convolutional layers, pooling layers, and fully connected layers. Set a suitable number and size of convolution kernels in the convolutional layer, and let it slide on the image to extract features such as vehicle outlines and license plate numbers. After the pooling layer, reduce the spatial dimension of the feature map and reduce the amount of calculation. Finally, integrate the features through the fully connected layer and output the key feature vector for subsequent analysis and judgment of traffic conditions. When using a recurrent neural network (RNN) to extract features from time series data such as traffic flow, the time series data is sequentially input into the RNN model. The model automatically learns to extract the correlation features between data at different time points and data change trends based on its internal neuron structure and parameters, providing a basis for analysis such as traffic flow prediction.

[0048] S140. Determine a data transmission strategy according to the network topology, data characteristics, and real-time network status of the edge device, and transmit the target feature data to the target device based on the data transmission strategy.

[0049] In this embodiment, the network topology refers to the network architecture or network layout that connects each edge device; it describes how devices are connected through the network, including the communication path between devices, the connection relationship between nodes, etc.; the topology structure will affect the efficiency and reliability of data transmission. The data transmission strategy refers to a decision-making plan designed for data transmission based on factors such as data characteristics, network conditions, and device topology; the transmission strategy may include selecting appropriate transmission protocols, paths, data caching strategies, frequencies, compression strategies, etc. to optimize the data transmission effect and ensure efficient and stable data transmission under different network conditions. The target device refers to the terminal device for data transmission, which may be a client device, server, or other device that receives and processes data; the target device will further process the received data according to application requirements.

[0050] Optionally, the data transmission strategy includes a transmission frequency strategy, a transmission path strategy, a transmission protocol strategy and a data cache strategy; the data characteristics include an integrity index, an accuracy index and a real-time index.

[0051] In this embodiment, the transmission frequency strategy refers to a strategy that determines how often data is transmitted; the frequency can be adjusted according to the real-time requirements of the application and the network conditions to ensure that data transmission is completed within a reasonable time interval. The transmission path strategy refers to a strategy that determines the path selection for data transmission from the source device to the target device; according to the network topology of the edge device and the real-time status of the network, the optimal transmission path is selected to improve transmission efficiency and reliability. The transmission protocol strategy refers to a strategy that determines the type of protocol used for data transmission; according to the characteristics of the data (such as integrity and accuracy requirements) and the real-time network conditions, a suitable transmission protocol is selected to ensure the correct transmission of data. The data cache strategy refers to a strategy for temporarily storing data during transmission to optimize transmission efficiency; according to the real-time network conditions, it is determined whether to use the cache, the size of the cache, and the update frequency of the cache to reduce network congestion or delay. The integrity index refers to the requirements for accuracy and completeness of data transmission. The accuracy index refers to the requirements for precision of data transmission. The real-time index refers to the requirements for timeliness and response speed of data transmission.

[0052] Furthermore, according to the network topology structure, data characteristics and real-time network conditions of the edge device, the data transmission strategy can be determined as follows: determining the transmission frequency strategy according to the real-time indicators in the data characteristics and the real-time network conditions; determining the transmission path strategy according to the network topology structure and the real-time network conditions of the edge device; determining the transmission protocol strategy according to the integrity indicators and accuracy indicators in the data characteristics, as well as the real-time network conditions; and determining the data caching strategy according to the real-time network conditions.

[0053] For example, for the transmission frequency strategy, from the perspective of network quality, when the network environment is good, such as low network latency, extremely low packet loss rate, and sufficient bandwidth, the edge device can appropriately increase the data transmission frequency to ensure that the data can be transmitted to the target device or server in a timely and efficient manner to meet the application scenario requirements with high data real-time requirements. For example, in intelligent traffic monitoring, if the network condition is good, the real-time position, speed and other key data of the vehicle can be transmitted at a high frequency, which is convenient for the background to quickly grasp the traffic dynamics and make accurate decisions. On the contrary, if the network quality is poor, such as high latency, high packet loss rate or limited bandwidth, if a high transmission frequency is still maintained at this time, it will only increase the risk of data loss and transmission errors, thereby affecting the reliability of the entire data transmission and subsequent data analysis and processing. Therefore, the edge device needs to reduce the transmission frequency accordingly, reduce the amount of data transmission in an unstable network environment, reduce the adverse effects of packet loss and delay, and gradually restore or increase the transmission frequency after the network quality improves. Considering the importance of data, for data that has extremely high real-time requirements and is related to key functions and decisions of the system, such as real-time vital signs data of patients in telemedicine and key equipment failure warning data in industrial automation, no matter how the network quality fluctuates, its relatively high transmission frequency should be guaranteed as much as possible, and the timely transmission of these important data should be prioritized. This can be achieved by allocating more network resources or adopting more reliable transmission strategies. For some data with low real-time requirements and relatively low importance, such as some auxiliary data of traffic flow statistics in intelligent traffic monitoring, its transmission frequency can be reduced to a greater extent when the network quality is poor, reducing the occupation of network bandwidth and allowing network resources to be more inclined to the transmission of important data. By comprehensively considering the network quality and data importance and dynamically changing the transmission frequency, edge devices can optimize the data transmission process in a complex and changing network environment, improve the overall data transmission efficiency and reliability, and meet the diverse needs of different application scenarios.

[0054] Optionally, determining the transmission path strategy based on the network topology and real-time network status of the edge device can be: determining a target path selection algorithm from candidate path selection algorithms based on the network topology and real-time network status of the edge device; and determining the transmission path strategy based on the target path selection algorithm.

[0055] In this embodiment, the candidate path selection algorithm may include the shortest path algorithm, the multi-path routing algorithm, the routing algorithm based on the quality of service, the machine learning algorithm, etc. The shortest path algorithm is based on the link weight (commonly, the delay is used as the weight) to calculate the shortest path from the source node to the target node; its principle is to gradually search and update the shortest distance of the adjacent nodes by continuously expanding the set of nodes with the shortest distance determined until the target node is found. The multi-path routing algorithm uses multiple paths for data transmission at the same time; when the network topology structure allows, by reasonably allocating data traffic to different paths, it can not only improve the reliability of data transmission, avoid data transmission interruption due to failure of a certain path, but also improve the transmission efficiency to a certain extent, and make full use of the bandwidth resources of multiple paths. The routing algorithm based on the quality of service refers to comprehensively considering various service quality indicators of the link, such as bandwidth, delay, packet loss rate, jitter, etc., and then selecting the optimal path that meets the conditions according to specific service quality requirements. The machine learning algorithm refers to training with the help of historical data and using the machine learning model to predict the optimal path; by analyzing the performance of each path in the past network at different time periods and different traffic conditions, learning the rules and characteristics therein, and then predicting the current and future optimal paths.

[0056] It is understandable that different transmission path selection algorithms have their own advantages and disadvantages. In actual edge device data transmission, it is often necessary to comprehensively consider multiple factors such as the specific network topology, application scenarios, and real-time network quality conditions, and flexibly select appropriate algorithms or combine multiple algorithms to achieve the optimal selection of data transmission paths.

[0057] Optionally, the transport protocol strategy includes at least one of the Transmission Control Protocol (TCP protocol), the User Datagram Protocol (UDP protocol), and the Quick UDP Internet Connection Protocol (QUIC protocol).

[0058] In this embodiment, TCP is a connection-oriented, reliable transmission protocol; before transmitting data, a connection needs to be established through a three-way handshake to ensure the reliability of communication between the two parties. It relies on mechanisms such as sequence numbers, confirmation responses, and timeout retransmissions to ensure the integrity and order of data. It also has congestion control and flow control mechanisms, which can effectively prevent network congestion and packet loss, and is suitable for application scenarios with high requirements for data reliability. UDP is a connectionless protocol. The sender and receiver do not need to establish a connection before communication. It is simple and efficient. It does not provide a confirmation or retransmission mechanism for data packets, nor does it guarantee the order of data packets. Therefore, it has high transmission efficiency and good real-time performance. It is suitable for application scenarios with high real-time requirements and can tolerate a small amount of data loss. The Fast UDP Internet Connection Protocol is a relatively new transmission protocol that aims to combine the advantages of TCP and UDP to provide better performance and security, especially in the case of poor network quality. It combines the low latency characteristics of UDP and a reliability mechanism similar to TCP. By implementing reliable transmission, congestion control, encryption and other functions at the application layer, it reduces the additional delay caused by the interaction between the network layer and the transport layer. It also supports fast connection establishment and multiplexing, and can transmit multiple data streams concurrently on the same connection, thereby improving the utilization of network resources.

[0059] For example, with regard to the transmission protocol strategy, if the network environment is relatively stable and the data has extremely high requirements for accuracy and integrity, such as financial transaction data transmission, key control instruction transmission in industrial automation, etc., the TCP protocol is given priority; if the transmitted data has high real-time requirements, such as real-time video surveillance screen transmission, online real-time game data interaction, etc., and a certain degree of data loss risk can be accepted, the UDP protocol is more suitable; and when the network environment is complex, changeable, and unstable and it is desired to take into account both transmission efficiency and reliability, the QUIC protocol can be used as the preferred solution. By adapting different transmission protocols, it can meet the data transmission needs in diverse edge computing application scenarios to the greatest extent.

[0060] For example, for data caching protocols, when the network has poor quality conditions such as high latency, high packet loss rate, or severe bandwidth shortage, if the edge device continues to forcibly transmit data to the target device or server, it is likely to cause a large amount of data loss, transmission errors, or a long period of waiting for transmission, affecting the operating efficiency of the entire system and the availability of data. At this time, the edge device can temporarily cache the data in local storage to avoid losses caused by blindly transmitting data under harsh network conditions.

[0061] Furthermore, during the transmission process, it is also necessary to follow certain rules and sequences, such as transmitting in sequence according to the timestamp order of the data or the priority of the data's importance, to ensure that the data can reach the target end completely and accurately, and can be combined with the previously mentioned strategies such as transmission frequency adjustment and transmission path selection to further optimize the transmission process of cached data and improve transmission efficiency and reliability. Through this local caching mechanism, edge devices can flexibly respond to fluctuations in network quality, ensure the continuity and stability of data transmission, improve the adaptability and fault tolerance of the entire data transmission system in complex network environments, and better serve various edge computing application scenarios.

[0062] The embodiment of the present application obtains the data to be transmitted by preprocessing the collected initial data, and monitors the network quality of the data transmission process in real time to obtain the real-time network status; compresses the data to be transmitted according to the data characteristics of the data to be transmitted and the real-time network status to obtain candidate data; extracts features from the candidate data according to the data characteristics to obtain target feature data; determines the data transmission strategy according to the network topology, data characteristics and real-time network status of the edge device, and transmits the target feature data to the target device based on the data transmission strategy. The above technical solution can reduce data transmission problems caused by unstable network environment and effectively improve the efficiency and reliability of data transmission by dynamically performing data compression, feature extraction and transmission strategy determination according to the real-time network status.

[0063] Embodiment 2

[0064] Figure 2 This is a flow chart of a data transmission method provided in accordance with the second embodiment of the present application. Based on the technical solutions of the above embodiments, this embodiment refines "compressing the data to be transmitted according to the data characteristics of the data to be transmitted and the real-time network status to obtain candidate data" into "determining the target compression method from the candidate compression methods according to the data characteristics of the data to be transmitted and the real-time network status; using the target compression method to compress the data to be transmitted to obtain candidate data". It should be noted that for the parts not described in detail in the embodiments of the present application, please refer to the relevant descriptions of other embodiments. Figure 2 As shown, the method includes:

[0065] S210: pre-process the collected initial data to obtain data to be transmitted, and monitor the network quality of the data transmission process in real time to obtain real-time network status.

[0066] S220: Determine a target compression method from candidate compression methods according to data characteristics of the data to be transmitted and real-time network conditions.

[0067] Optionally, the candidate compression methods include lossless compression and approximate compression; the data characteristics include real-time indicators.

[0068] Furthermore, if the real-time index of the data to be transmitted meets the high real-time requirement and the real-time network condition is good, lossless compression is determined as the target compression method; if the real-time index of the data to be transmitted meets the low real-time requirement, approximate compression is determined as the target compression method.

[0069] For example, data is selectively compressed based on the importance of the data and the quality of the network. For example, in an intelligent traffic monitoring system, for data with high real-time requirements (such as vehicle speed), a lossless compression algorithm is used to ensure the integrity of the data so that traffic conditions can be accurately analyzed later; while for data with low real-time requirements (such as vehicle type), an approximate compression algorithm can be used to obtain a higher compression rate, reduce the amount of data transmitted over the network, and reduce the burden of network transmission.

[0070] S230: Compress the data to be transmitted using a target compression method to obtain candidate data.

[0071] In an optional implementation, if the target compression method is lossless compression, a Huffman tree is constructed based on the character frequencies of candidate characters in the data to be transmitted; based on the structure of the Huffman tree, corresponding character codes are generated for the candidate characters, and the candidate characters are replaced with the character codes to obtain candidate data.

[0072] In this embodiment, the Huffman tree is a binary tree for data compression, which is constructed based on the frequency of occurrence of characters and is used to generate the optimal binary code; in Huffman coding, characters with higher frequencies correspond to shorter codes, while characters with lower frequencies correspond to longer codes, thereby reducing the overall code length. Character frequency refers to the number of times or proportions of each character in a given text; it is used to measure the frequency of occurrence of each character in the text. Character encoding refers to the process of mapping characters to digital codes that can be understood by computers; it assigns a unique binary code to each character; the purpose of character encoding is to ensure that text data can be correctly stored and transmitted between computer systems.

[0073] In another optional implementation, if the target compression method is approximate compression, the data to be transmitted is divided into at least one data vector block, and a clustering algorithm is used to analyze the at least one data vector block to obtain at least one codeword; for each data vector block, the data vector block is matched with at least one codeword, and the successfully matched codeword replaces the data vector block to obtain candidate data.

[0074] In this embodiment, the data vector block is to divide the data to be compressed into several small blocks, each of which contains a certain number of data elements (such as numbers, pixel values, etc.) in order to process or compress it; in approximate compression, the data is divided into multiple data vector blocks, and each vector block can be regarded as a separate processing unit. Clustering algorithm refers to an unsupervised learning method used to divide the elements in a data set into multiple groups or clusters according to a certain similarity measure; common clustering algorithms include K-means algorithm, hierarchical clustering algorithm, etc.; clustering algorithm is used to analyze data blocks, identify data patterns and features, and thus generate a more representative "codeword" to replace the original data block. Codeword refers to a unique identifier representing specific data in the process of data compression or encoding, which is used to represent the compressed substitute of the data vector block.

[0075] For example, when using Huffman coding for lossless compression, the frequency of each character in the data to be compressed must first be counted, and the corresponding Huffman tree must be constructed. Then, the corresponding code is generated for each character according to the tree structure, and then the characters in the original data are replaced with the code for compressed storage. When using vector quantization for approximate compression, it is necessary to first determine the appropriate codebook generation method, divide the vector blocks according to the characteristics of the data, and then use the vectors in the codebook to approximate the actual data blocks to complete the compression process.

[0076] S240: Extract features from the candidate data according to data characteristics to obtain target feature data.

[0077] S250. Determine a data transmission strategy according to the network topology, data characteristics, and real-time network status of the edge device, and transmit the target feature data to the target device based on the data transmission strategy.

[0078] In an optional implementation manner, after the target device receives the target feature data, it is necessary to decompress the target feature data to obtain the decompressed target feature data.

[0079] Specifically, the target device usually sets up a corresponding decompression module, which encapsulates decompression algorithms corresponding to various data compression algorithm types, and sets an associated algorithm model to make the compression algorithm correspond to the decompression algorithm, so as to realize the decompression operation of data compressed by any compression algorithm. For example, first identify the algorithm identification information carried in the compressed data header to determine which compression algorithm is used, and then call the corresponding decompression algorithm for decompression processing to ensure that the data can smoothly enter the subsequent feature reconstruction and data analysis.

[0080] It should be noted that the decompression algorithm may include at least one of a lossless compression algorithm and an approximate compression algorithm, and the embodiment of the present application does not specifically limit this.

[0081] Furthermore, if the decompressed data contains feature data after feature extraction, feature reconstruction and further data analysis are required to obtain the final displayable data.

[0082] Specifically, in terms of feature reconstruction, for data that uses machine learning methods to extract features, such as features extracted based on support vector machines (SVM), decision trees, random forests and other algorithms, the reconstruction must be based on the reverse operation of the rules and parameters based on the corresponding algorithms in the feature extraction stage; for example, SVM extracts features by mapping data to a high-dimensional space to find the optimal hyperplane, so feature reconstruction may require restoring the extracted features to the original data space according to the mapping relationship, but this process may be more complicated, and because some information may be abstracted during feature extraction, completely accurate reconstruction is difficult; feature reconstruction extracted by decision trees and random forests can be based on the node division rules in the tree structure and feature importance ranking information, The features can be restored to a form closer to the original data representation, which is convenient for subsequent intuitive understanding and analysis. The reconstruction of features extracted by deep learning methods such as convolutional neural networks (CNN) and recurrent neural networks (RNN) also has its own characteristics. When reconstructing image features extracted by CNN, deconvolution, upsampling and other operations are performed to gradually restore the feature maps processed by the convolution layer and pooling layer to the original image size and feature form. For example, in intelligent traffic monitoring, the extracted vehicle outlines, license plate numbers and other features are reconstructed to restore image information that can be manually viewed or further analyzed. The reconstruction of time series features extracted by RNN is based on its memory and processing mechanism in the time dimension, and reverse deduction according to the sequence order to restore the relevant features of the original time series data. For example, time series data such as traffic flow can be reconstructed in order to more accurately analyze the changing trends of traffic conditions. The feature reconstruction of data extracted based on domain knowledge is mainly based on specific methods and rules designed in advance; for example, for the features such as color, texture, shape extracted from image data, the reconstruction should be combined with the knowledge of the image field and related algorithms to recombin and render these features to generate an intuitive image; for the statistical features such as mean, variance, autocorrelation coefficient extracted from time series data, the appearance of the original time series data should be restored as much as possible through corresponding mathematical inverse operations and data supplementation, so as to provide a complete data basis for subsequent analysis. After completing the feature reconstruction, data analysis should be carried out; in the intelligent traffic monitoring scenario, the speed change trend of vehicles, the peak and trough periods of traffic flow, the congestion of different sections of roads, etc. can be analyzed, and the decision-making basis for traffic command and dispatch can be provided through statistical analysis and correlation analysis of a large amount of data; in the smart home scenario, the changing laws of environmental data are analyzed, combined with the characteristics of user behavior patterns, intelligent temperature adjustment, electrical control and other functions are realized, and the comfort and convenience of home life are improved; in the industrial automation scenario, the characteristics of machine operation data are analyzed, abnormal conditions are discovered in time, fault diagnosis and predictive maintenance are carried out, and the stable and efficient operation of the production process is guaranteed.Different application scenarios will adopt different data analysis methods and models, such as descriptive statistical analysis to understand the basic characteristics of the data, cluster analysis to perform user grouping or device status classification, time series analysis to predict future trends, etc. These methods are used in combination to explore the value behind the data and realize the practical functions of the present invention in various scenarios.

[0083] Furthermore, the displayable data after data analysis can be displayed to meet the needs of users to view and use data in different application scenarios.

[0084] For example, in the intelligent traffic monitoring scenario, data display can be presented in the form of a large visual screen; for example, in the traffic command center, the large screen can display real-time traffic flow maps, using different colors and icons to identify the congestion level of different sections, vehicle speed and other information, so that managers can intuitively see the overall traffic conditions; it can also display monitoring video images of various intersections, as well as traffic flow trend charts generated based on data analysis, statistics of accident-prone areas, and other content, to facilitate traffic scheduling decisions. At the same time, it supports the data drill function. By clicking on a specific section on the map or a data point in the chart, you can further view more detailed vehicle information, historical traffic data, etc. on that section, which facilitates in-depth analysis. In the smart home scenario, data display can be performed on mobile phone applications or home smart terminal devices; users can view the temperature and humidity curves of each room in the room to understand the environmental comfort; view the status of doors and windows, the operating status of electrical equipment, and other information, and can display recent energy consumption in the form of charts or simple status indicators, etc., to assist users in reasonably controlling the use of electrical appliances and saving energy; in addition, it can also display intelligent reminders and suggestions based on user behavior pattern characteristics, such as reminding whether the air conditioning temperature needs to be adjusted based on the user's habitual sleep time and indoor temperature preference. In terms of industrial automation, the real-time operating parameters of each production equipment, such as speed, temperature, pressure and other data, can be displayed on the large screen in the factory's monitoring room, and whether the data is within the normal range can be intuitively presented in the form of dashboards, bar charts, etc.; display equipment fault warning information, and inform maintenance personnel to handle it in a timely manner in a striking color or pop-up window prompt; at the same time, present the historical trend chart of equipment operation data, which is convenient for analyzing the performance changes of the equipment and formulating preventive maintenance plans; in addition, it supports the linkage display of data, such as clicking on the fault information of a certain equipment, which can automatically associate and display the maintenance records and parts replacement history of the equipment, etc., to help maintenance personnel quickly locate and solve problems.

[0085] It should be noted that the data display function usually relies on professional data visualization tools or self-developed display platforms. These tools and platforms have rich chart type support, flexible layout settings, and convenient data interaction functions. They can display processed data in a customized manner according to different application scenarios and user needs, clearly present the value behind the data to users, and meet diverse practical application needs.

[0086] The embodiment of the present application obtains the data to be transmitted by preprocessing the collected initial data, and monitors the network quality of the data transmission process in real time to obtain the real-time network status; determines the target compression method from the candidate compression methods according to the data characteristics of the data to be transmitted and the real-time network status; compresses the data to be transmitted using the target compression method to obtain the candidate data; extracts features from the candidate data according to the data characteristics to obtain the target feature data; determines the data transmission strategy according to the network topology, data characteristics and real-time network status of the edge device, and transmits the target feature data to the target device based on the data transmission strategy. The above technical scheme can reduce data transmission problems caused by unstable network environment and effectively improve the efficiency and reliability of data transmission by dynamically performing data compression, feature extraction and transmission strategy determination according to the real-time network status.

[0087] Embodiment 3

[0088] Figure 3 This is a schematic diagram of the structure of a data transmission device provided according to the third embodiment of the present application, which can be applied to edge devices for network data transmission. The data transmission device can be implemented in the form of hardware and / or software, and the data transmission device can be configured in a computer device, such as an edge device. Figure 3 As shown, the device comprises:

[0089] The data acquisition module 310 is used to pre-process the collected initial data to obtain the data to be transmitted, and to monitor the network quality of the data transmission process in real time to obtain the real-time network status;

[0090] The data compression module 320 is used to compress the data to be transmitted according to the data characteristics of the data to be transmitted and the real-time network status to obtain candidate data;

[0091] The feature extraction module 330 is used to extract features from the candidate data according to the data characteristics to obtain target feature data;

[0092] The strategy determination module 340 is used to determine the data transmission strategy according to the network topology, data characteristics and real-time network status of the edge device, and transmit the target feature data to the target device based on the data transmission strategy.

[0093] The embodiment of the present application obtains the data to be transmitted by preprocessing the collected initial data, and monitors the network quality of the data transmission process in real time to obtain the real-time network status; compresses the data to be transmitted according to the data characteristics of the data to be transmitted and the real-time network status to obtain candidate data; extracts features from the candidate data according to the data characteristics to obtain target feature data; determines the data transmission strategy according to the network topology, data characteristics and real-time network status of the edge device, and transmits the target feature data to the target device based on the data transmission strategy. The above technical solution can reduce data transmission problems caused by unstable network environment and effectively improve the efficiency and reliability of data transmission by dynamically performing data compression, feature extraction and transmission strategy determination according to the real-time network status.

[0094] Optionally, the data compression module 320 includes:

[0095] A method determination unit, configured to determine a target compression method from candidate compression methods according to data characteristics of the data to be transmitted and real-time network conditions;

[0096] The data compression unit is used to compress the data to be transmitted by adopting a target compression method to obtain candidate data.

[0097] Optionally, the candidate compression methods include lossless compression and approximate compression; the data characteristics include a real-time index; and accordingly, the method determination unit is specifically used to:

[0098] If the real-time index of the data to be transmitted meets the high real-time requirement and the real-time network condition is good, lossless compression is determined as the target compression method;

[0099] If the real-time index of the data to be transmitted meets the low real-time requirement, approximate compression is determined as the target compression method.

[0100] Optionally, a data compression unit is used to:

[0101] If the target compression method is lossless compression, a Huffman tree is constructed according to the character frequencies of the candidate characters in the data to be transmitted; according to the structure of the Huffman tree, corresponding character codes are generated for the candidate characters, and the candidate characters are replaced with the character codes to obtain candidate data;

[0102] If the target compression method is approximate compression, the data to be transmitted is divided into at least one data vector block, and a clustering algorithm is used to analyze the at least one data vector block to obtain at least one codeword; for each data vector block, the data vector block is matched with at least one codeword, and the successfully matched codeword replaces the data vector block to obtain candidate data.

[0103] Optionally, the data characteristic includes a data type; accordingly, the feature extraction module 330 is specifically used to:

[0104] According to the data type in the data characteristics, determine the target feature extraction method from the candidate feature extraction methods; the candidate feature extraction methods include machine learning algorithms and deep learning algorithms; the deep learning algorithms include convolutional neural networks and recurrent neural networks;

[0105] The target feature extraction method is used to extract features from the candidate data to obtain target feature data.

[0106] Optionally, the data transmission strategy includes a transmission frequency strategy, a transmission path strategy, a transmission protocol strategy and a data cache strategy; the data characteristics include an integrity index, an accuracy index and a real-time index; accordingly, the strategy determination module 340 is specifically used to:

[0107] Determine the transmission frequency strategy based on the real-time indicators in the data characteristics and the real-time network status;

[0108] Determine the transmission path strategy based on the network topology and real-time network status of edge devices;

[0109] Determine the transmission protocol strategy based on the integrity and accuracy indicators in the data characteristics and the real-time network status;

[0110] Determine the data caching strategy based on the real-time network status.

[0111] The data transmission device provided in the embodiments of the present application can execute the data transmission method provided in any embodiment of the present application, and has the corresponding functional modules and beneficial effects for executing each data transmission method.

[0112] According to an embodiment of the present application, the present application also provides an electronic device, a readable storage medium and a computer program product.

[0113] Embodiment 4

[0114] Figure 4 4 is a schematic diagram of the structure of an electronic device 410 that implements the data transmission method of an embodiment of the present application. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.

[0115] like Figure 4 As shown, the electronic device 410 includes at least one processor 411, and a memory connected to the at least one processor 411 in communication, such as a read-only memory (ROM) 412, a random access memory (RAM) 413, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 411 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 412 or the computer program loaded from the storage unit 418 to the random access memory (RAM) 413. In RAM413, various programs and data required for the operation of the electronic device 410 can also be stored. The processor 411, ROM412 and RAM413 are connected to each other via a bus 414. An input / output (I / O) interface 415 is also connected to the bus 414.

[0116] Multiple components in the electronic device 410 are connected to the I / O interface 415, including: an input unit 416, such as a keyboard, a mouse, etc.; an output unit 417, such as various types of displays, speakers, etc.; a storage unit 418, such as a disk, an optical disk, etc.; and a communication unit 419, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 419 allows the electronic device 410 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0117] The processor 411 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 411 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 411 performs the various methods and processes described above, such as a data transmission method.

[0118] In some embodiments, the data transmission method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 418. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 410 via the ROM 412 and / or the communication unit 419. When the computer program is loaded into the RAM 413 and executed by the processor 411, one or more steps of the data transmission method described above may be performed. Alternatively, in other embodiments, the processor 411 may be configured as the data transmission method in any other appropriate manner (e.g., by means of firmware).

[0119] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0120] Computer programs for implementing the methods of the present application may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data transmission device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0121] In the context of the present application, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. 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.

[0122] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0123] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by 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), a blockchain network, and the Internet.

[0124] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0125] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution of this application can be achieved, and this document is not limited here.

[0126] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A data transmission method, characterized in that: Applied to an edge device, the method comprises: Pre-process the collected initial data to obtain the data to be transmitted, and monitor the network quality of the data transmission process in real time to obtain the real-time network status; According to the data characteristics of the data to be transmitted and the real-time network status, compressing the data to be transmitted to obtain candidate data; According to the data characteristics, feature extraction is performed on the candidate data to obtain target feature data; A data transmission strategy is determined according to the network topology of the edge device, the data characteristics and the real-time network status, and based on the data transmission strategy, the target feature data is transmitted to the target device.

2. The method according to claim 1, characterized in that According to the data characteristics of the data to be transmitted and the real-time network status, the data to be transmitted is compressed to obtain candidate data, including: Determining a target compression method from candidate compression methods according to data characteristics of the data to be transmitted and the real-time network status; The target compression method is adopted to compress the data to be transmitted to obtain candidate data.

3. The method according to claim 2, characterized in that The candidate compression methods include lossless compression and approximate compression; the data characteristics include a real-time index; accordingly, determining a target compression method from the candidate compression methods according to the data characteristics of the data to be transmitted and the real-time network status includes: If the real-time index of the data to be transmitted meets the high real-time requirement and the real-time network condition is good, the lossless compression is determined as the target compression method; If the real-time index of the data to be transmitted meets the low real-time requirement, the approximate compression is determined as the target compression method.

4. The method according to claim 3, characterized in that The adopting the target compression method to compress the data to be transmitted to obtain candidate data includes: If the target compression method is lossless compression, a Huffman tree is constructed according to the character frequencies of the candidate characters in the data to be transmitted; a corresponding character code is generated for the candidate characters according to the structure of the Huffman tree, and the candidate characters are replaced with the character code to obtain candidate data; If the target compression method is approximate compression, the data to be transmitted is divided into at least one data vector block, and a clustering algorithm is used to analyze the at least one data vector block to obtain at least one codeword; for each data vector block, the data vector block is matched with the at least one codeword, and the successfully matched codeword replaces the data vector block to obtain candidate data.

5. The method according to claim 1, characterized in that The data characteristics include data types; accordingly, extracting features from the candidate data according to the data characteristics to obtain target feature data includes: Determine a target feature extraction method from candidate feature extraction methods according to the data type in the data characteristics; the candidate feature extraction methods include machine learning algorithms and deep learning algorithms; the deep learning algorithms include convolutional neural networks and recurrent neural networks; The target feature extraction method is used to extract features from the candidate data to obtain target feature data.

6. The method according to claim 1, characterized in that The data transmission strategy includes a transmission frequency strategy, a transmission path strategy, a transmission protocol strategy and a data cache strategy; the data characteristics include an integrity index, an accuracy index and a real-time index; accordingly, according to the network topology of the edge device, the data characteristics and the real-time network status, the data transmission strategy is determined, including: Determining a transmission frequency strategy according to the real-time index in the data characteristics and the real-time network status; Determine a transmission path strategy according to the network topology of the edge device and the real-time network status; Determining the transmission protocol strategy according to the integrity index and the accuracy index in the data characteristics and the real-time network status; The data caching strategy is determined according to the real-time network status.

7. A data transmission device, characterized in that: Configured on an edge device, the device includes: The data acquisition module is used to pre-process the collected initial data to obtain the data to be transmitted, and to monitor the network quality of the data transmission process in real time to obtain the real-time network status; A data compression module, used for compressing the data to be transmitted according to the data characteristics of the data to be transmitted and the real-time network status to obtain candidate data; A feature extraction module, used to extract features from the candidate data according to the data characteristics to obtain target feature data; The strategy determination module is used to determine the data transmission strategy according to the network topology of the edge device, the data characteristics and the real-time network status, and transmit the target feature data to the target device based on the data transmission strategy.

8. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the data transmission method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the data transmission method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the data transmission method according to any one of claims 1 to 6.

Images