A remote desktop control system and method based on cloud computing

Abstract

The present invention discloses a remote desktop control system and method based on cloud computing, which relates to the technical field of digital information transmission. The remote desktop control system includes: a cloud computing platform module, a data analysis module, a performance optimization module, and a remote control strategy module. The present invention can extract key factors affecting remote desktop performance from interactive feature data, improve the ability to accurately identify performance bottlenecks, help optimize resource allocation and performance optimization strategies, and improve the efficiency of overall remote desktop control. It can also predict future remote desktop control performance and accurately predict possible performance change trends, thereby providing a scientific basis for adjusting resource allocation and optimization strategies in advance, helping to reduce latency and improve user experience. At the same time, based on performance change trends and the user's real-time interactive status, the system can dynamically adjust the remote desktop control strategy to achieve intelligent optimization of the transmission protocol.

Description

Technical Field

[0001] The present invention relates to the technical field of digital information transmission, and in particular to a remote desktop control system and method based on cloud computing. Background Art

[0002] With the rapid development of science and technology, automated intelligent control has gradually become an indispensable part of people's lives. Remote desktop control is a concrete manifestation of this technology. Remote desktop control technology enables device operation through a remote connection, allowing various tasks to be completed without direct contact with the device screen, greatly improving work efficiency and convenience. This technology has been widely used in fields such as office, education, and healthcare. Especially with the increasing popularity of remote work and online education, remote desktop control has become an important means to improve resource utilization efficiency and conserve human resources.

[0003] In the existing technology, it is not convenient to extract the key factors affecting remote desktop performance from the interactive feature data, improve the ability to accurately identify performance bottlenecks, optimize resource allocation and performance optimization strategies, improve the efficiency of overall remote desktop control, and predict future remote desktop control performance. It is impossible to accurately predict possible performance change trends, thereby providing a scientific basis for adjusting resource allocation and optimization strategies in advance, increasing latency and reducing user experience.

[0004] Currently, no effective solutions have been proposed for the problems in related technologies. Summary of the Invention

[0005] In response to the problems in the related art, the present invention proposes a remote desktop control system and method based on cloud computing to overcome the above technical problems existing in the existing related art.

[0006] To this end, the specific technical solutions adopted in the present invention are as follows:

[0007] According to one aspect of the present invention, a remote desktop control system based on cloud computing is provided, the remote desktop control system comprising:

[0008] The cloud computing platform module is used to obtain real-time interaction data of remote desktop control based on the cloud computing platform, and process the real-time interaction data to obtain interaction feature data of remote desktop control;

[0009] A data analysis module is used to analyze the obtained interactive feature data of remote desktop control using a factor analysis algorithm to identify key factors affecting the performance of remote desktop control;

[0010] The performance optimization module is used to establish a performance optimization model based on key factors, and use the performance optimization model to predict the remote desktop control performance at future moments to obtain the change trend of the remote desktop control performance;

[0011] The remote control policy module is used to dynamically adjust the control policy of the remote desktop and optimize the transmission protocol based on the changing trend and the real-time interaction status of the user.

[0012] Furthermore, the cloud computing platform module includes:

[0013] The data acquisition module is used to establish a secure connection between the client and the remote host by using a remote communication protocol and to collect real-time interactive data controlled by the remote desktop;

[0014] The data processing module is used to extract features from the collected real-time interaction data and generate interaction feature data for remote desktop control.

[0015] Furthermore, using the factor analysis algorithm, we analyzed the interactive feature data of remote desktop control and identified the key factors that affect the performance of remote desktop control, including:

[0016] Randomly generate an initial interactive feature dataset, calculate the performance impact objective function value of each interactive feature, hierarchically sort the interactive features and calculate the distribution density to form an optimal interactive feature dataset;

[0017] Using the interactive feature screening algorithm, the factor features that have the greatest impact on performance are selected from the preferred feature set, and the factor features are copied and expanded to form a new interactive feature dataset;

[0018] Reorganize and adjust the new interactive feature data set, remove the factor features that do not meet the performance constraints, and use the initialization function to regenerate the interactive feature data that meets the constraints;

[0019] The primary interaction feature dataset is merged with the secondary interaction feature dataset to form a new optimal interaction feature dataset, and duplicate features are removed to finally generate a non-duplicate optimal interaction feature dataset;

[0020] Determine whether the number of iterations reaches a preset threshold. If so, stop the iteration and output the key factors affecting the remote desktop control performance. Otherwise, continue the iteration.

[0021] Furthermore, using the interactive feature screening algorithm, we select the factor features that have the greatest impact on performance from the optimal feature set, copy and expand the factor features, and form a new interactive feature dataset including:

[0022] Collect the optimal interactive feature data of remote desktop control, use the correlation analysis method to evaluate the correlation between features, and generate a subset of interactive feature data classified by performance impact;

[0023] Randomly select several sample training sets from the interaction feature data subset, calculate the gain value of each interaction feature's impact on performance, build a performance analysis model, and screen key interaction features;

[0024] Eliminate incomplete interactive feature data, verify the data with the highest integrity, and divide it into several feature data subsets as training sets;

[0025] The performance analysis model is used to train the key feature data subsets, calculate their impact probabilities under different performance states, generate performance feature templates, screen out the subsets with the best performance, and form a new interactive feature data set.

[0026] Furthermore, the performance optimization module includes:

[0027] A model building module for building a performance optimization model based on the identified key factors affecting remote desktop control performance;

[0028] The performance prediction module is used to predict the remote desktop control performance at future moments by using the established performance optimization model to obtain the change trend of the remote desktop control performance.

[0029] Furthermore, based on the identified key factors affecting remote desktop control performance, a performance optimization model is constructed, including:

[0030] Obtain a set of key factors that affect remote desktop control performance, and divide the key factor set into a training set and a test set;

[0031] Taking any key factor as the main reference feature, combined with other key factors, they are input into the ensemble learning model to construct multiple performance optimization sub-models;

[0032] The performance of all performance optimization sub-models is evaluated on the training set, and the best-performing performance optimization sub-model is selected for fusion to form the final performance optimization model.

[0033] Furthermore, the established performance optimization model is used to predict the remote desktop control performance in the future, and the change trends of the remote desktop control performance are obtained, including:

[0034] Input key factors affecting remote desktop control performance into the performance optimization model as basic data for future performance prediction;

[0035] Using the performance optimization model to calculate the performance indicators at future times, a preliminary prediction result of remote desktop control performance is generated;

[0036] Based on the prediction results, the future trend of remote desktop control performance is analyzed, the performance fluctuation area is identified, and based on the performance fluctuation area, the trend of remote desktop control performance is obtained.

[0037] According to another aspect of the present invention, a remote desktop control method based on cloud computing is provided, the remote desktop control method comprising the following steps:

[0038] S1. Based on the cloud computing platform, real-time interaction data of remote desktop control is obtained, and the real-time interaction data is processed to obtain interaction feature data of remote desktop control;

[0039] S2. Analyze the obtained interactive feature data of remote desktop control using a factor analysis algorithm to identify key factors affecting remote desktop control performance;

[0040] S3. Based on the key factors, a performance optimization model is established, and the performance optimization model is used to predict the remote desktop control performance at future moments to obtain the change trend of the remote desktop control performance;

[0041] S4. Based on the changing trend and the real-time interaction status of the user, the control strategy of the remote desktop is dynamically adjusted to optimize the transmission protocol.

[0042] The beneficial effects of the present invention are:

[0043] 1. The present invention can extract key factors affecting remote desktop performance from interactive feature data, improve the ability to accurately identify performance bottlenecks, help optimize resource allocation and performance optimization strategies, and improve the efficiency of overall remote desktop control. It can also predict future remote desktop control performance and accurately predict possible performance change trends, thereby providing a scientific basis for adjusting resource allocation and optimization strategies in advance, helping to reduce latency and improve user experience. At the same time, based on performance change trends and real-time user interaction status, the system can dynamically adjust the remote desktop control strategy to achieve intelligent optimization of the transmission protocol. By optimizing the transmission protocol, data transmission latency can be further reduced, network bandwidth utilization can be improved, and the fluency and reliability of remote desktop control can be enhanced.

[0044] 2. The present invention accurately identifies the key factors affecting remote desktop performance through a factor analysis algorithm, optimizes the feature screening and data reorganization process, utilizes the optimal feature data set, eliminates redundant features and dynamically adjusts the data structure, thereby improving the validity of feature data and the prediction accuracy of the model, thereby improving the performance stability, resource utilization and cross-platform compatibility of remote desktop control, and providing users with a more efficient and smooth remote operation experience.

[0045] 3. The present invention uses a performance optimization model to accurately predict future remote desktop control performance. By inputting key factors and calculating the model, it generates performance change trends and identifies fluctuation areas. It can predict performance issues in advance and provide a scientific basis for optimizing resource allocation and dynamic policy adjustments, thereby meeting the needs of efficient remote operation in complex network environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 This is a principle block diagram of a remote desktop control system based on cloud computing according to an embodiment of the present invention;

[0048] Figure 2 The present invention is a flowchart of a remote desktop control method based on cloud computing according to an embodiment of the present invention.

[0049] In the picture:

[0050] 1. Cloud computing platform module; 2. Data analysis module; 3. Performance optimization module; 4. Remote control strategy module. DETAILED DESCRIPTION

[0051] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention. They are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. By referring to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and the advantages of the present invention.

[0052] According to an embodiment of the present invention, a remote desktop control system and method based on cloud computing are provided.

[0053] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to an embodiment of the present invention, a remote desktop control system based on cloud computing includes:

[0054] The cloud computing platform module 1 is used to obtain real-time interaction data of remote desktop control based on the cloud computing platform, and process the real-time interaction data to obtain interaction feature data of remote desktop control;

[0055] It's important to note that lightweight data collection modules (agents) are deployed on user terminals and remote desktop devices to capture real-time interactions between users and the remote desktop, as well as device status. These modules upload the collected data to the cloud platform. This data is sent to the cloud computing platform via secure and efficient transmission protocols such as WebSocket, HTTP / 2, or gRPC. WebSocket is particularly well-suited for low-latency data streaming, ensuring real-time upload of interactive data.

[0056] Specifically, real-time interaction data includes:

[0057] 1) User operation data: records of user operations on the remote desktop, such as mouse movement, clicks, dragging, and keyboard input.

[0058] 2) Screen update data: changes in the remote desktop screen, such as window movement, image refresh, video playback, etc.

[0059] 3) Network transmission data: real-time transmission performance indicators such as packet delay, packet loss rate, bandwidth usage, and network jitter.

[0060] 4) System performance data: CPU, memory, GPU usage, load, temperature and other hardware performance information of the remote desktop device.

[0061] 5) Connection status data: real-time status of session connection, including connection time, disconnection times, recovery time, etc.

[0062] 7) User environment data: basic environment data such as the user terminal’s operating system, browser information, and device model.

[0063] Specifically, the interactive feature data includes:

[0064] 1) Operation response time characteristics: the delay from user operation to remote desktop response.

[0065] 2) Screen refresh characteristics: screen update rate, frame rate, resolution and other characteristics related to visual smoothness.

[0066] 3) Network performance characteristics: bandwidth utilization, latency fluctuation, packet loss rate, and other characteristics that affect the smoothness of interaction.

[0067] 4) Resource usage characteristics: CPU, memory, and GPU resource usage of the remote desktop and user terminal.

[0068] 5) Interaction stability characteristics: stability of data packet transmission, session duration, error recovery rate, etc.

[0069] 6) User behavior characteristics: frequency of user operations, repetitive behaviors, and usage ratios of different functions.

[0070] Data analysis module 2, used to analyze the obtained interactive feature data of remote desktop control using a factor analysis algorithm to identify key factors affecting the performance of remote desktop control;

[0071] Specifically, the key factors include:

[0072] 1) Network performance factors

[0073] Bandwidth: The amount of available network bandwidth directly affects the data transfer rate.

[0074] Latency: The time interval between user operation and remote desktop response is an important factor affecting the smoothness of interaction.

[0075] Packet loss rate: The ratio of data packets lost during transmission, which may cause image freezes or operation failures.

[0076] Jitter: The degree of fluctuation in network delay, which affects the stability of real-time data.

[0077] 2) Remote desktop system performance factors

[0078] Server CPU / memory usage: Whether the system resources are sufficient determines the response speed.

[0079] GPU processing power: For graphics-intensive operations such as video playback or 3D rendering, GPU performance is crucial.

[0080] Disk I / O performance: The speed of the storage during file transfers or read / write operations can affect task completion time.

[0081] 3) User terminal performance factors

[0082] Terminal device performance: The CPU, memory, and GPU capabilities of the user device will affect the decoding speed and data rendering performance.

[0083] Screen resolution: Data transfer and rendering requirements are higher at high resolutions.

[0084] Operating system and browser: The compatibility of different operating systems and browsers may affect performance.

[0085] 4) Interactive behavioral factors

[0086] Operation frequency: The frequency of user operations (such as the intensity of mouse clicks and keyboard input) affects the system load.

[0087] Interaction mode: Whether it is high-frequency input (such as typing) or complex operations (such as dragging and dropping files) has different impacts on performance.

[0088] 5) Image and video quality factors

[0089] Compression algorithm: The compression rate of image and video data is inversely proportional to transmission efficiency and visual quality.

[0090] Refresh rate: The screen refresh rate has a significant impact on the smoothness of video playback and dynamic operations.

[0091] Color depth: High color depth will increase the amount of data and affect the transmission speed.

[0092] 6) Security protocol factors

[0093] Encryption algorithm: The higher the encryption strength, the better the data security, but it may increase calculation and transmission delays.

[0094] Authentication method: Complex authentication may increase interaction initialization time.

[0095] Performance optimization module 3 is used to establish a performance optimization model based on key factors, and use the performance optimization model to predict the remote desktop control performance at future moments to obtain the change trend of the remote desktop control performance;

[0096] The remote control strategy module 4 is used to dynamically adjust the control strategy of the remote desktop and optimize the transmission protocol based on the change trend and the real-time interaction status of the user.

[0097] It needs to be explained that the performance change trend of the remote desktop is analyzed in real time, and key performance fluctuation areas (such as increased latency, insufficient bandwidth, etc.) are captured as the basis for optimization; user operation frequency, input type (mouse, keyboard, etc.) and task complexity (file transfer, video playback, etc.) are collected to determine the interactive load characteristics; according to performance changes and user interaction needs, the image compression algorithm, resolution and color depth are dynamically adjusted to reduce the amount of data to improve transmission efficiency; the transmission protocol parameters are dynamically switched, such as adjusting the WebSocket buffer size, data blocking strategy, and accelerated packet retransmission mechanism to reduce the impact of latency and packet loss; the allocation of CPU, GPU and memory resources is optimized in the cloud to ensure that high-load tasks (such as video rendering) receive priority support and improve overall performance; according to real-time network bandwidth and latency fluctuations, the data transmission rate is adaptively adjusted to ensure the continuity and smoothness of the user experience; the effectiveness of the optimized strategy is regularly evaluated, and the control strategy is further adjusted through feedback loops to ensure that the transmission protocol always adapts to the user and network environment.

[0098] In this optional embodiment, the cloud computing platform module 1 includes:

[0099] The data acquisition module is used to establish a secure connection between the client and the remote host by using a remote communication protocol and to collect real-time interactive data controlled by the remote desktop;

[0100] The data processing module is used to extract features from the collected real-time interaction data and generate interaction feature data for remote desktop control.

[0101] In this optional embodiment, feature extraction is performed on the collected real-time interaction data to generate interaction feature data of remote desktop control, including:

[0102] Obtain duplicate data from the remote desktop's real-time interaction data, perform denoising, filtering, and smoothing, and concatenate unprocessed interaction data rows to generate a complete interaction data table.

[0103] Based on accurate interaction data sets, we construct interaction feature trends and extract the change feature indicators of remote desktop interaction status by calculating the difference in the change amplitude of feature trends at adjacent moments.

[0104] A preset change threshold is set. If the change difference exceeds the threshold, a new key interaction moment set is added; otherwise, the existing key interaction moment set is retained.

[0105] Check the distribution of key interaction moments. If the distribution is too concentrated, integrate adjacent key interaction moments; otherwise, retain the current key moment set as the key feature data of remote desktop interaction.

[0106] In this optional embodiment, the obtained interactive feature data of remote desktop control is analyzed using a factor analysis algorithm to identify key factors affecting the performance of remote desktop control, including:

[0107] Randomly generate an initial interactive feature dataset, calculate the performance impact objective function value of each interactive feature, hierarchically sort the interactive features and calculate the distribution density to form an optimal interactive feature dataset;

[0108] Using the interactive feature screening algorithm, the factor features that have the greatest impact on performance are selected from the preferred feature set, and the factor features are copied and expanded to form a new interactive feature dataset;

[0109] Reorganize and adjust the new interactive feature data set, remove the factor features that do not meet the performance constraints, and use the initialization function to regenerate the interactive feature data that meets the constraints;

[0110] The primary interaction feature dataset is merged with the secondary interaction feature dataset to form a new optimal interaction feature dataset, and duplicate features are removed to finally generate a non-duplicate optimal interaction feature dataset;

[0111] Determine whether the number of iterations reaches a preset threshold. If so, stop the iteration and output the key factors affecting the remote desktop control performance. Otherwise, continue the iteration.

[0112] Specifically, the factor analysis algorithm is a modified NSGA-II algorithm. NSGA-II is a classic multi-objective optimization algorithm that effectively finds a solution set that satisfies multiple objectives when used to solve multi-objective optimization problems. This paper improves on the basic NSGA-II algorithm by designing an efficient integer encoding method and introducing a deduplication operation.

[0113] In this optional embodiment, an interactive feature screening algorithm is used to select the factor features that have the greatest impact on performance from the preferred feature set, and the factor features are copied and expanded to form a new interactive feature data set including:

[0114] Collect the optimal interactive feature data of remote desktop control, use the correlation analysis method to evaluate the correlation between features, and generate a subset of interactive feature data classified by performance impact;

[0115] Randomly select several sample training sets from the interaction feature data subset, calculate the gain value of each interaction feature on performance, build a performance analysis model (i.e., decision tree classifier) ​​and screen key interaction features;

[0116] Eliminate incomplete interactive feature data, verify the data with the highest integrity, and divide it into several feature data subsets as training sets;

[0117] The performance analysis model is used to train the key feature data subsets, calculate their impact probabilities under different performance states, generate performance feature templates, screen out the subsets with the best performance, and form a new interactive feature data set.

[0118] Specifically, the interactive feature screening algorithm is an association rule algorithm, which is an algorithm used to mine the association relationships between items in a data set. Through the association rule algorithm, key association relationships can be discovered from massive interactive feature data, and the feature combinations that have the greatest impact on performance can be identified, providing strong support for building and optimizing performance analysis models.

[0119] In this optional embodiment, the formula for calculating the gain value of each interactive feature's impact on performance is:

[0120] ;

[0121] Where Z(T, A) represents the gain value of the interaction feature A to the performance target T; P(ti) represents the probability of the i-th global state appearing in the performance target T; It represents the probability of the jth conditional state in the performance target T appearing when the value of the interaction feature A is v; P(v) represents the probability that the value of feature A is v; n represents the total number of global states; m represents the total number of conditional states.

[0122] In this optional embodiment, the performance optimization module 3 includes:

[0123] A model building module for building a performance optimization model based on the identified key factors affecting remote desktop control performance;

[0124] The performance prediction module is used to predict the remote desktop control performance at future moments by using the established performance optimization model to obtain the change trend of the remote desktop control performance.

[0125] In this optional embodiment, building a performance optimization model based on the identified key factors affecting remote desktop control performance includes:

[0126] Obtain a set of key factors that affect remote desktop control performance, and divide the key factor set into a training set and a test set;

[0127] Taking any key factor as the main reference feature, combined with other key factors, they are input into the ensemble learning model to construct multiple performance optimization sub-models;

[0128] The performance of all performance optimization sub-models is evaluated on the training set, and the best-performing performance optimization sub-model is selected for fusion to form the final performance optimization model.

[0129] In this optional embodiment, the established performance optimization model is used to predict the remote desktop control performance at future times, and the obtained change trend of the remote desktop control performance includes:

[0130] Input key factors affecting remote desktop control performance into the performance optimization model as basic data for future performance prediction;

[0131] Using the performance optimization model to calculate the performance indicators at future times, a preliminary prediction result of remote desktop control performance is generated;

[0132] Based on the prediction results, the future trend of remote desktop control performance is analyzed, the performance fluctuation area is identified, and based on the performance fluctuation area, the trend of remote desktop control performance is obtained.

[0133] It is important to clarify that, first, key factors influencing remote desktop control performance are identified. These factors may include network bandwidth, latency, packet loss rate, user operation frequency, screen refresh rate, and terminal device performance. These key factors are then fed into a performance optimization model as input parameters, forming the foundational data for predictive analysis. Based on these key factors, the performance optimization model calculates future performance indicators using established association rules and prediction algorithms (such as regression analysis and machine learning models), generating preliminary predictions. These performance indicators may include core performance parameters such as response time, data transfer rate, screen refresh rate, and resource utilization. Based on the predictions generated by the performance optimization model, the performance trends of remote desktop control are analyzed. By combining performance indicators at different time points, potential future performance fluctuations are identified. These fluctuations may correspond to conditions such as network congestion, frequent user operations, or insufficient system resources. Once these fluctuations are identified, a deeper analysis of the trends is conducted to further clarify the scope and potential duration of the fluctuations. For example, when network bandwidth fluctuates, the system predicts its specific impact on screen refresh rate and response time, as well as the likely recovery time. Performance trends are output as data reports or visualizations, helping operations personnel or the system's automatic optimization module understand performance changes in real time. Furthermore, based on the prediction results, it recommends adjusting control strategies, such as optimizing image compression, reducing refresh rates, or adjusting transmission protocol parameters, to minimize the impact of fluctuations on the user experience.

[0134] According to another embodiment of the present invention, Figure 2 As shown, a remote desktop control method based on cloud computing is also provided, and the remote desktop control method includes the following steps:

[0135] S1. Based on the cloud computing platform, real-time interaction data of remote desktop control is obtained, and the real-time interaction data is processed to obtain interaction feature data of remote desktop control;

[0136] S2. Analyze the obtained interactive feature data of remote desktop control using a factor analysis algorithm to identify key factors affecting remote desktop control performance;

[0137] S3. Based on the key factors, a performance optimization model is established, and the performance optimization model is used to predict the remote desktop control performance at future moments to obtain the change trend of the remote desktop control performance;

[0138] S4. Based on the changing trend and the real-time interaction status of the user, the control strategy of the remote desktop is dynamically adjusted to optimize the transmission protocol.

[0139] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A remote desktop control system based on cloud computing, characterized in that: The remote desktop control system includes: The cloud computing platform module is used to obtain real-time interaction data of remote desktop control based on the cloud computing platform, and process the real-time interaction data to obtain interaction feature data of remote desktop control; A data analysis module is used to analyze the obtained interactive feature data of remote desktop control using a factor S analysis algorithm to identify key factors affecting the performance of remote desktop control; The performance optimization module is used to establish a performance optimization model based on key factors, and use the performance optimization model to predict the remote desktop control performance at future moments to obtain the change trend of the remote desktop control performance; Remote control strategy module, which is used to dynamically adjust the remote desktop control strategy and optimize the transmission protocol based on the changing trend and the real-time user interaction status; The factor analysis algorithm is used to analyze the obtained interactive feature data of the remote desktop control to identify key factors affecting the performance of the remote desktop control, including: Randomly generate an initial interactive feature dataset, calculate the performance impact objective function value of each interactive feature, hierarchically sort the interactive features and calculate the distribution density to form an optimal interactive feature dataset; Using the interactive feature screening algorithm, the factor features that have the greatest impact on performance are selected from the preferred feature set, and the factor features are copied and expanded to form a new interactive feature dataset; Reorganize and adjust the new interactive feature data set, remove the factor features that do not meet the performance constraints, and use the initialization function to regenerate the interactive feature data that meets the constraints; The primary interaction feature dataset is merged with the secondary interaction feature dataset to form a new optimal interaction feature dataset, and duplicate features are removed to finally generate a non-duplicate optimal interaction feature dataset; Determine whether the number of iterations reaches a preset threshold. If so, stop the iteration and output the key factors affecting the remote desktop control performance. Otherwise, continue the iteration. The interactive feature screening algorithm is used to select the factor features that have the greatest impact on performance from the preferred feature set, and the factor features are copied and expanded to form a new interactive feature data set, including: Collect the optimal interactive feature data of remote desktop control, use the correlation analysis method to evaluate the correlation between features, and generate a subset of interactive feature data classified by performance impact; Randomly select several sample training sets from the interaction feature data subset, calculate the gain value of each interaction feature's impact on performance, build a performance analysis model, and screen key interaction features; Eliminate incomplete interactive feature data, verify the data with the highest integrity, and divide it into several feature data subsets as training sets; Use the performance analysis model to train the key feature data subsets, calculate their impact probabilities under different performance states, and generate performance feature templates to screen out the subsets with the best performance to form a new interactive feature dataset; The formula for calculating the gain value of each interactive feature's impact on performance is: ； Where, Z ( T , A ) represents the interaction feature A Performance goals T The gain value of P ( t i ) represents the performance target T Middle i The probability of a global state occurring; Indicated in the interaction feature A The value of v When the performance target T Middle j The probability of a conditional state occurring; P ( v ) indicates characteristics A The value of v probability; n Represents the total number of global states; m Indicates the total number of condition states.

2. A remote desktop control system based on cloud computing according to claim 1, characterized in that: The cloud computing platform module includes: The data acquisition module is used to establish a secure connection between the client and the remote host by using a remote communication protocol and to collect real-time interactive data controlled by the remote desktop; The data processing module is used to extract features from the collected real-time interaction data and generate interaction feature data for remote desktop control.

3. A remote desktop control system based on cloud computing according to claim 2, characterized in that: The feature extraction of the collected real-time interaction data to generate interaction feature data of remote desktop control includes: Obtain duplicate data from the remote desktop's real-time interaction data, perform denoising, filtering, and smoothing, and concatenate unprocessed interaction data rows to generate a complete interaction data table. Based on accurate interaction data sets, we construct interaction feature trends and extract the change feature indicators of remote desktop interaction status by calculating the difference in the change amplitude of feature trends at adjacent moments. A preset change threshold is set. If the change difference exceeds the threshold, a new key interaction moment set is added; otherwise, the existing key interaction moment set is retained. Check the distribution of key interaction moments. If the distribution is too concentrated, integrate adjacent key interaction moments; otherwise, retain the current key moment set as the key feature data of remote desktop interaction.

4. The remote desktop control system based on cloud computing according to claim 1, characterized in that: The performance optimization module includes: A model building module for building a performance optimization model based on the identified key factors affecting remote desktop control performance; The performance prediction module is used to predict the remote desktop control performance at future moments by using the established performance optimization model to obtain the change trend of the remote desktop control performance.

5. The remote desktop control system based on cloud computing according to claim 4, characterized in that: The construction of a performance optimization model based on the identified key factors affecting remote desktop control performance includes: Obtain a set of key factors that affect remote desktop control performance, and divide the key factor set into a training set and a test set; Taking any key factor as the main reference feature, combined with other key factors, they are input into the ensemble learning model to construct multiple performance optimization sub-models; The performance of all performance optimization sub-models is evaluated on the training set, and the best-performing performance optimization sub-model is selected for fusion to form the final performance optimization model.

6. The remote desktop control system based on cloud computing according to claim 5, characterized in that: The remote desktop control performance is predicted by using the established performance optimization model in the future, and the change trend of the remote desktop control performance is obtained, including: Input key factors affecting remote desktop control performance into the performance optimization model as basic data for future performance prediction; Using the performance optimization model to calculate the performance indicators at future times, a preliminary prediction result of remote desktop control performance is generated; Based on the prediction results, the future trend of remote desktop control performance is analyzed, the performance fluctuation area is identified, and based on the performance fluctuation area, the trend of remote desktop control performance is obtained.

7. A remote desktop control method based on cloud computing, characterized in that: The remote desktop control method includes the following steps: S1. Based on the cloud computing platform, real-time interaction data of remote desktop control is obtained, and the real-time interaction data is processed to obtain interaction feature data of remote desktop control; S2. Analyze the obtained interactive feature data of remote desktop control using a factor analysis algorithm to identify key factors affecting remote desktop control performance, including the following steps: Randomly generate an initial interactive feature dataset, calculate the performance impact objective function value of each interactive feature, hierarchically sort the interactive features and calculate the distribution density to form an optimal interactive feature dataset; Using the interactive feature screening algorithm, the factor features that have the greatest impact on performance are selected from the preferred feature set, and the factor features are copied and expanded to form a new interactive feature dataset; Reorganize and adjust the new interactive feature data set, remove the factor features that do not meet the performance constraints, and use the initialization function to regenerate the interactive feature data that meets the constraints; The primary interaction feature dataset is merged with the secondary interaction feature dataset to form a new optimal interaction feature dataset, and duplicate features are removed to finally generate a non-duplicate optimal interaction feature dataset; Determine whether the number of iterations reaches a preset threshold. If so, stop the iteration and output the key factors affecting the remote desktop control performance. Otherwise, continue the iteration. The interactive feature screening algorithm is used to select the factor features that have the greatest impact on performance from the preferred feature set, and the factor features are copied and expanded to form a new interactive feature data set, including: Collect the optimal interactive feature data of remote desktop control, use the correlation analysis method to evaluate the correlation between features, and generate a subset of interactive feature data classified by performance impact; Randomly select several sample training sets from the interaction feature data subset, calculate the gain value of each interaction feature's impact on performance, build a performance analysis model, and screen key interaction features; Eliminate incomplete interactive feature data, verify the data with the highest integrity, and divide it into several feature data subsets as training sets; Use the performance analysis model to train the key feature data subsets, calculate their impact probabilities under different performance states, and generate performance feature templates to screen out the subsets with the best performance to form a new interactive feature dataset; The formula for calculating the gain value of each interactive feature's impact on performance is: ； Where, Z ( T , A ) represents the interaction feature A Performance goals T The gain value of P ( t i ) represents the performance target T Middle i The probability of a global state occurring; Indicated in the interaction feature A The value of v When the performance target T Middle j The probability of a conditional state occurring; P ( v ) indicates characteristics A The value of v probability; n Represents the total number of global states; m Indicates the total number of condition states; S3. Based on the key factors, a performance optimization model is established, and the performance optimization model is used to predict the remote desktop control performance at future moments to obtain the change trend of the remote desktop control performance; S4. Based on the changing trend and the real-time interaction status of the user, the control strategy of the remote desktop is dynamically adjusted to optimize the transmission protocol.

Images