Data security protection method based on ownership separation

Abstract

The invention relates to the technical field of data security processing, in particular to a data security protection method based on ownership separation, which comprises the following steps: S1, setting the uploading time, uploading mode and uploading position of private feature data and non-private feature data; s2, respectively uploading the data to an uploading position for storage; s3, setting and optimizing the uploading time, the uploading mode and the uploading position of the private feature data; s4, the uploading time, the uploading mode and the uploading position of the private feature data are adjusted in an optimization mode for setting the uploading time, the uploading mode and the uploading position; s5, deleting the private data of the abnormal edge node; s6, optimizing the judgment process of the storage state of the private data of the abnormal edge node; and S7, adjusting the uploading time, the uploading mode and the uploading position of the private data of the abnormal edge node in the step S1. The data security protection efficiency of the artificial intelligence robot is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data security processing, and in particular, to a data security protection method based on separation of ownership rights. Background Art

[0002] In the field of artificial intelligence robots, since there is a lot of contact with user privacy, it is crucial to implement the separation of ownership rights of user privacy data. This protection measure can make users feel safer and more at ease when using artificial intelligence robots to handle delicate tasks in life and work. However, current artificial intelligence robots still face many challenges in user data processing and it is difficult to effectively balance the conflict between computing power and data encryption. In addition, although federated learning helps to protect privacy, its time cost and operation cost are relatively high; while centralized training cannot fully guarantee user privacy and data security. Therefore, these problems need to be solved urgently to promote the further development of artificial intelligence robots.

[0003] Chinese Patent Publication No.: CN117376015A discloses a security authentication method for home robots, which is applied to the field of robot security and aims at the security problems of privacy protection and data security faced by home robots; in the robot stage, identity registration is first completed in the server, and after the identity registration process is completed, relevant information such as the ID and public key of the robot node are stored on the server; relevant information such as the ID and public key of the server are also stored in the robot node for subsequent identity verification and key exchange; after mutual authentication, the server and the robot node establish a session key using the disclosed information to ensure the security of communication. However, this solution is difficult to effectively balance the conflict between computing power and data encryption. Summary of the Invention

[0004] Therefore, the present invention provides a data security protection method based on separation of ownership rights to overcome the problem of low efficiency of data security protection of artificial intelligence robots caused by the high time cost and operation cost of federated training in the prior art, while centralized training cannot ensure user privacy and there is a conflict between device computing power and encryption processing.

[0005] To achieve the above object, the present invention provides a data security protection method based on separation of ownership rights, including: Step S1, identifying the running data according to the data privacy recognition algorithm to obtain private feature data and non-private feature data, and setting the upload time, upload method and upload location of the private feature data and the non-private feature data; Step S2, uploading the private feature data and the non-private feature data to the upload location for storage respectively according to the upload time and upload method of the private feature data and the non-private feature data; Step S3, judge the attributes of the private feature data according to the generated position of the private feature data, obtain the edge node private data and the central node private data, and optimize the setting of the upload time, upload method and upload location of the private feature data according to the attributes of the private feature data; Step S4, judge the abnormal situation of the edge node private data, obtain the abnormal edge node private data and the normal edge node private data, and adjust the optimization method of setting the upload time, upload method and upload location of the private feature data; Step S5, judge the storage status of the abnormal edge node private data according to the call times and storage duration of the abnormal edge node private data, and delete the abnormal edge node private data according to the storage status of the abnormal edge node private data; Step S6, optimize the judgment process of the storage status of the abnormal edge node private data according to the storage occupancy ratio of the adjacent edge nodes; Step S7, judge the importance of the abnormal edge node private data according to the secondary call rate of the abnormal edge node private data, and adjust the upload time, upload method and upload location of the abnormal edge node private data in Step S1 according to the importance of the abnormal edge node private data.

[0006] Furthermore, the data security protection method based on ownership separation is applied to the data processing system of an artificial intelligence robot. There are edge nodes with a preset number of processes and 1 central node in the artificial intelligence robot. Each edge node is connected to two adjacent edge nodes, and each edge node is also connected to the central node.

[0007] Furthermore, in Step S1, identify the running data according to the data privacy recognition algorithm to obtain the private feature data and non-private feature data, and set the upload time, upload method and upload location of the running data, where: When the running data is private feature data, set the upload time of the running data as the pre-fuzzy time, the upload method of the running data as the pre-fuzzy upload, and the upload location of the running data as the central node and adjacent edge nodes; When the running data is non-private feature data, set the upload time of the running data as the current time, the upload method of the running data as non-fuzzy upload, and the upload location of the running data as the central node.

[0008] Furthermore, in Step S2, upload the running data to the upload location for storage according to the upload time and upload method of the running data through the asymmetric encryption method to achieve the ownership separation of the running data.

[0009] Further, in the step S3, the attributes of the private feature data are judged according to the generation location of the private feature data to obtain edge node private data and central node private data, and the upload time, upload method, and upload location of the private feature data are optimized according to the attributes of the private feature data, where: When the attribute of the private feature data is edge node private data, the upload time of the private feature data is not to upload, the upload method of the private feature data is not to perform fuzzy upload, and the upload location of the private feature data is no upload location; When the attribute of the private feature data is central node private data, the upload time, upload method, and upload location of the private feature data are not optimized.

[0010] Further, in the step S4, the abnormal conditions of the edge node private data are judged to obtain abnormal edge node private data and normal edge node private data, and the optimized methods for setting the upload time, upload method, and upload location of the private feature data are adjusted. The abnormal edge node private data is judged as non-closed-loop reincarnation data, and pre-fuzzy processing upload is performed, and the upload location is adjusted to be backed up and encrypted to the adjacent edge node to realize non-uploaded federated training.

[0011] Further, in the step S5, the storage status of the abnormal edge node private data is judged according to the call times and storage duration of the abnormal edge node private data, and the abnormal edge node private data is deleted according to the storage status of the abnormal edge node private data.

[0012] Further, in the step S6, the judgment process of the storage status of the abnormal edge node private data is optimized according to the storage occupancy ratio of the adjacent edge node, and the call times and storage duration are reduced.

[0013] Further, in the step S7, the importance of the abnormal edge node private data is judged according to the secondary call rate of the abnormal edge node private data, and the upload time, upload method, and upload location of the abnormal edge node private data in the step S1 are adjusted. If the secondary call rate is low, it is adjusted to post-fuzzy, and if the secondary call rate is high, the adjustment location is central storage.

[0014] Further, in the step S7, if the secondary call rate is high, the adjustment location is central storage. Description of the Drawings

[0015] Figure 1 It is a flowchart of the data security protection method based on ownership separation in this embodiment. Detailed Embodiment

[0016] To make the objectives and advantages of the present invention more clear and understandable, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0017] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0018] Please refer to Figure 1 as shown, which is a schematic flowchart of the data security protection method based on ownership separation in this embodiment, including: Step S1, identify the running data according to the data privacy recognition algorithm to obtain private feature data and non-private feature data, and set the upload time, upload method, and upload location of the private feature data and non-private feature data; Step S2, upload the private feature data and non-private feature data to the upload location for storage respectively according to the upload time and upload method of the private feature data and non-private feature data; Step S3, judge the attributes of the private feature data according to the location where the private feature data is generated to obtain edge node private data and central node private data, and optimize the setting of the upload time, upload method, and upload location of the private feature data according to the private feature data attributes; Step S4, judge the abnormal conditions of the edge node private data to obtain abnormal edge node private data and normal edge node private data, and adjust the optimization method for setting the upload time, upload method, and upload location of the private feature data; Step S5, judge the storage status of the abnormal edge node private data according to the call times and storage duration of the abnormal edge node private data, and delete the abnormal edge node private data according to the storage status of the abnormal edge node private data; Step S6, optimize the judgment process of the storage status of the abnormal edge node private data according to the storage occupancy ratio of adjacent edge nodes; Step S7, judge the importance of the abnormal edge node private data according to the secondary call rate of the abnormal edge node private data, and adjust the upload time, upload method, and upload location of the abnormal edge node private data in Step S1 according to the importance of the abnormal edge node private data.

[0019] Specifically, this method is applied to the data processing system of an artificial intelligence robot to balance the settings of federated training and centralized training in the artificial intelligence robot, so as to solve the conflict between device computing power and encryption processing, thereby improving the data security protection efficiency of the artificial intelligence robot.

[0020] Specifically, the data security protection method based on ownership separation is applied to the data processing system of an AI robot. The AI robot is provided with edge nodes with a preset number of processes and 1 central node. Each edge node is connected to two adjacent edge nodes, and each edge node is also connected to the central node.

[0021] Specifically, in the step S1, the running data is identified according to the data privacy recognition algorithm to obtain private feature data, non-private feature data, and the upload time, upload method, and upload location of the running data are set, where: When the running data is private feature data, the upload time of the running data is set as the pre-fuzzy time, the upload method of the running data is set as pre-fuzzy upload, and the upload location of the running data is the central node and adjacent edge nodes; When the running data is non-private feature data, the upload time of the running data is set as the current time, the upload method of the running data is set as non-fuzzy upload, and the upload location of the running data is the central node.

[0022] Specifically, in the step S2, the running data is uploaded to the upload location for storage according to the upload time and upload method of the running data by the asymmetric encryption method to achieve the separation of ownership of the running data.

[0023] Specifically, in the step S3, the attributes of the private feature data are judged according to the generation location of the private feature data to obtain edge node private data and central node private data, and the upload time, upload method, and upload location of the private feature data are optimized according to the attributes of the private feature data, where: When the attribute of the private feature data is edge node private data, the upload time of the private feature data is set as not uploading, the upload method of the private feature data is set as not performing fuzzy upload, and the upload location of the private feature data is no upload location; When the attribute of the private feature data is central node private data, the upload time, upload method, and upload location of the private feature data are not optimized.

[0024] Specifically, in the step S4, the abnormal conditions of the edge node private data are judged to obtain abnormal edge node private data and normal edge node private data, and the optimization method of setting the upload time, upload method, and upload location of the private feature data is adjusted. The abnormal edge node private data is judged as non-closed-loop reincarnation data and pre-fuzzy processed for upload, and the upload location is adjusted to be encrypted and backed up to adjacent edge nodes to achieve non-uploaded federated training.

[0025] Specifically, in the step S5, the storage status of the private data of the abnormal edge node is judged according to the call times and storage duration of the private data of the abnormal edge node, and the private data of the abnormal edge node is deleted according to the storage status of the private data of the abnormal edge node.

[0026] Specifically, in the step S6, the judgment process of the storage status of the private data of the abnormal edge node is optimized according to the storage occupancy ratio of the adjacent edge nodes, and the call times and storage duration are reduced.

[0027] Specifically, in the step S7, the importance of the private data of the abnormal edge node is judged according to the secondary call rate of the private data of the abnormal edge node, and the upload time, upload method and upload location of the private data of the abnormal edge node in the step S1 are adjusted according to the importance of the private data of the abnormal edge node. If the secondary call rate is low, it is adjusted to post-positioned fuzzy storage; if the secondary call rate is high, it is adjusted to central storage.

[0028] Specifically, in the step S7, if the secondary call rate is high, it is adjusted to central storage.

[0029] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A data security protection method based on ownership separation, characterized in that: include: Step S1, identifying the operation data according to the data privacy identification algorithm, obtaining private feature data and non-private feature data, and setting the upload time, upload method and upload location of the private feature data and the non-private feature data; Step S2, uploading the private feature data and the non-private feature data to an upload location for storage according to the upload time and upload method of the private feature data and the non-private feature data; Step S3, judging the attributes of the private feature data according to the location where the private feature data is generated, obtaining edge node private data and central node private data, and optimizing the upload time, upload method and upload location of the private feature data according to the attributes of the private feature data; Step S4, judging the abnormal situation of edge node private data, obtaining abnormal edge node private data and normal edge node private data, and adjusting the optimization method for setting the upload time, upload method and upload location of the private feature data; Step S5, judging the storage status of the abnormal edge node private data according to the number of calls and storage duration of the abnormal edge node private data, and deleting the abnormal edge node private data according to the storage status of the abnormal edge node private data; Step S6, optimizing the judgment process of the storage status of the private data of the abnormal edge node according to the storage ratio of the adjacent edge nodes; Step S7, judging the importance of the private data of the abnormal edge nodes according to the secondary call rate of the private data of the abnormal edge nodes, and adjusting the upload time, upload method and upload location of the private data of the abnormal edge nodes in step S1 according to the importance of the private data of the abnormal edge nodes.

2. The data security protection method based on ownership separation according to claim 1 is characterized in that: The data security protection method based on separation of ownership is applied to the data processing system of an artificial intelligence robot, wherein the artificial intelligence robot is provided with a preset number of edge nodes and one central node, each edge node is connected to two adjacent edge nodes, and each edge node is connected to the central node.

3. The data security protection method based on ownership separation according to claim 1 is characterized in that: In step S1, the operation data is identified according to the data privacy identification algorithm to obtain private feature data and non-private feature data, and the upload time, upload method and upload location of the operation data are set, wherein: When the operation data is private feature data, the upload time of the operation data is set to the pre-fuzzy time, the upload method of the operation data is pre-fuzzy upload, and the upload location of the operation data is the central node and the adjacent edge node; When the operation data is non-private feature data, the upload time of the operation data is set to the current time, the upload method of the operation data is set to non-fuzzy upload, and the upload location of the operation data is set to the central node.

4. The data security protection method based on ownership separation according to claim 1 is characterized in that: In the step S2, the operation data is uploaded to the upload location for storage according to the upload time and upload method of the operation data by an asymmetric encryption method, so as to realize the separation of ownership of the operation data.

5. The data security protection method based on ownership separation according to claim 1 is characterized in that: In step S3, the private feature data attributes are judged according to the location where the private feature data is generated, and the edge node private data and the central node private data are obtained. The upload time, upload method and upload location of the private feature data are optimized according to the private feature data attributes, wherein: When the attribute of the private feature data is edge node private data, the upload time of the private feature data is set to not upload, the upload method of the private feature data is set to not fuzzy upload, and the upload location of the private feature data is set to no upload location; When the private feature data attribute is the central node private data, the upload time, upload method and upload location of the private feature data are not optimized.

6. The data security protection method based on ownership separation according to claim 1 is characterized in that: In step S4, the abnormal situation of the private data of the edge node is judged to obtain the private data of the abnormal edge node and the private data of the normal edge node, and the optimization method for setting the upload time, upload method and upload location of the private feature data is adjusted, the private data of the abnormal edge node is judged as non-closed-loop reincarnation data, and is uploaded with pre-fuzzy processing, and the upload location is adjusted to backup and encrypt to the adjacent edge node to realize federated training without uploading.

7. The data security protection method based on ownership separation according to claim 1 is characterized in that: In the step S5, the storage status of the abnormal edge node private data is judged according to the number of calls and storage duration of the abnormal edge node private data, and the abnormal edge node private data is deleted according to the storage status of the abnormal edge node private data.

8. The data security protection method based on ownership separation according to claim 1 is characterized in that: In step S6, the process of determining the storage status of private data of abnormal edge nodes is optimized according to the storage ratio of adjacent edge nodes, thereby reducing the number of calls and storage time.

9. The data security protection method based on ownership separation according to claim 1 is characterized in that: In step S7, the importance of the private data of the abnormal edge node is judged according to the secondary call rate of the private data of the abnormal edge node, and the upload time, upload method and upload position of the private data of the abnormal edge node in step S1 are adjusted according to the importance of the private data of the abnormal edge node. If the secondary call rate is low, it is adjusted to post-fuzzy; if the secondary call rate is high, the position is adjusted to central storage.

10. The data security protection method based on ownership separation according to claim 1 is characterized in that: In step S7, the secondary call rate is high and the position is adjusted to be stored in the center.

Images