Space target collaborative credible sensing method based on digital encryption

Abstract

The invention provides a space target collaborative credible sensing method based on digital encryption, belongs to the technical field of artificial intelligence, and solves the problem that the practical application of a multi-spacecraft collaborative sensing system is restricted due to the fact that a traditional data transmission method between a detection spacecraft and a decision spacecraft is difficult to balance data security, transmission efficiency and calculation cost. Comprising the steps of 1, encrypting detection image information of a detection sensor carried by a detection spacecraft through a digital encryption technology, and transmitting an encrypted image and an encrypted public key to a decision-making spacecraft through a public channel; 2, the decision spacecraft carries out model training by taking the received encrypted image and the encrypted public key as input and taking a variable L corresponding to a task as a supervised quantity, and outputs a variable # imgabs0 # required by the task; 3, the optimal observation position of the detection spacecraft is determined by combining a decision system and the variable # imgabs1 #, a planning instruction is transmitted to the detection spacecraft after being encrypted, and the optimal observation position of the detection spacecraft is determined by the decision system and the variable # imgabs1 #; and multi-spacecraft collaborative perception optimization is realized.

Description

Technical Field

[0001] The invention relates to a space target collaborative trusted perception method based on digital encryption, belonging to the technical field of spacecraft transmission services. Background Art

[0002] With the rapid development of space exploration and national defense science and technology, multi-spacecraft collaborative operations have become an important strategy to enhance space situational awareness capabilities. Under the guidance of the national strategy of "space power", my country's aerospace technology is accelerating towards a highly intelligent and networked direction. Multi-spacecraft collaborative systems are not only an important direction for national aerospace science and technology innovation, but also a key way to enhance space situational awareness and defense capabilities. Faced with the complex international space competition environment, building a safe and efficient spacecraft collaborative perception system has become an urgent need for the development of national aerospace science and technology.

[0003] However, the existing multi-spacecraft collaborative perception methods have serious defects in data transmission security. In the traditional communication mode, data transmission between detection spacecraft and decision-making spacecraft is extremely vulnerable to eavesdropping and tampering. Especially in open channels, the enemy may obtain key mission information by intercepting and parsing detection images, which seriously threatens the operational security of spacecraft. Existing encryption methods either have high computational complexity and poor real-time performance, or have insufficient encryption strength and cannot effectively protect the transmission of sensitive information between spacecraft. Traditional methods find it difficult to balance data security, transmission efficiency and computing cost, which restricts the practical application of multi-spacecraft collaborative perception systems. Summary of the invention

[0004] In order to solve the problem that traditional data transmission methods between detection spacecraft and decision-making spacecraft are difficult to balance data security, transmission efficiency and computing cost, which restricts the practical application of multi-spacecraft collaborative perception systems, the present invention proposes a collaborative trusted perception method for space targets based on digital encryption.

[0005] The technical solution adopted by the present invention to solve the above-mentioned problem is: the present invention comprises the following steps:

[0006] Step 1: Encrypt the detection image information of the detection sensor carried by the detection spacecraft through digital encryption technology, and transmit the encrypted image and encryption public key to the decision-making spacecraft through an open channel;

[0007] Step 2: The decision-making spacecraft uses the received encrypted image and encrypted public key as input, uses the variable L corresponding to the task as the supervision quantity to train the model, and outputs the variable required by the task.

[0008] Step 3: Combine decision systems and variables Confirm the optimal observation position of the exploration spacecraft, encrypt the planning instructions and transmit them to the exploration spacecraft to achieve collaborative perception optimization of multiple spacecraft.

[0009] Preferably, step 1 specifically includes:

[0010] Step 1.1: Use a randomly generated AES symmetric key K AES The detection data in the image information I is scrambled, forward diffused and reverse diffused to generate the encrypted image information C I =E AES (I,K AES );

[0011] Step 1.2: Use RSA public key For the AES symmetric key K AES Encrypt and generate a ciphertext symmetric key

[0012]

[0013] Step 1.3: Transmit the encrypted image information C through a public channel I To the decision-making spacecraft, and use the error correction code ECC and digital signature DS to the image information C I The transmission process is enhanced to generate transmission data

[0014]

[0015] Step 1.4: Pass the ciphertext symmetric key privacy transmission to decision-making spacecraft;

[0016] Step 1.5: Regularly change the AES symmetric key K AES and RSA key pair in, is the RSA public key, To determine the spacecraft's own RSA private key, and to change the AES symmetric key K according to the set key replacement period T AES and RSA key pair to update.

[0017] Preferably, step 2 specifically includes:

[0018] Step 2.1: Use the decision-making spacecraft's own RSA private key Symmetric key for ciphertext Decrypt and recover the AES symmetric key

[0019] Step 2.2: Use the recovered AES symmetric key For transmission data Decrypt and get the original image information I = DAES (C I ,K AES ) where the variable L includes structural semantics and relative posture;

[0020] Step 2.3: Use the decrypted image information I and the variable L corresponding to the task as the labeled data D = (I, L) to train the Transformer perception model;

[0021] Step 2.4: The trained Transformer perception model outputs the variables required by the task Among them, the variable Including the target structure semantics and relative posture required by the task.

[0022] Preferably, step 2.3 specifically includes:

[0023] In the training process of the Transformer perception model, the image information I is used as the model input, the task variable L is used as the supervision variable, the generalization ability of the Transformer perception model is improved through data enhancement technology, and the Adam optimizer is used to minimize the loss function L to train the Transformer perception model;

[0024] The update formula of Adam optimizer is:

[0025]

[0026] In formulas (1) and (2), β 1 is the decay rate of the first-order moment estimate, is the gradient calculated at time step t, v t is the second-order moment estimate at time t, is the bias-corrected estimate of the second-order moment at time t, v t-1 is the second-order moment estimate at time t-1, β 2 is the decay rate of the second-order moment estimate, m t is the first-order moment estimate at time t, is the bias-corrected estimate of the first-order moment at time t, m t-1 is the first-order moment estimate at time t-1, is the decay rate of the first-order moment estimate raised to the tth power, is the t-th power of the decay rate of the second-order moment estimate, α is the initial learning rate, ε is a small constant to avoid dividing by 0 when updating parameters, and θ t is the parameter vector at time t, θ t-1 is the parameter vector at time t-1.

[0027] Preferably, step 3 specifically includes:

[0028] Step 3.1: Define the implicit expression y=f(x) between the observation position x of the detection spacecraft and the target observation result y, and use the reinforcement learning model to fit the implicit expression y to obtain the output result

[0029] Step 3.2: Collect a training dataset for the decision model where x i and i are the sequences representing different “observation positions-observation results” obtained through the test in step 2, and the loss function is used To train the decision model based on reinforcement learning, the Adam optimizer is used to update the parameter θ;

[0030] Step 3.3: Using optimization strategies and variables Get the optimal observation position x * Let the objective function J(x) = -g(x,θ * ) minimize;

[0031] Step 3.4: Use the AES symmetric encryption algorithm to generate a random AES symmetric key K AES , for the optimal observation position x * Encrypt and generate encrypted instruction C x =E AES (x * ,K AES );

[0032] Step 3.5: Use RSA public key For the AES symmetric key K AES Encrypt and generate a ciphertext symmetric key

[0033]

[0034] Step 3.6: Transmit the encrypted command C through the public channel x and the ciphertext symmetric key

[0035] Step 3.7: Decision spacecraft uses its own RSA private key Symmetric key for ciphertext Decrypt and recover the AES symmetric key And use the recovered AES symmetric key K AES For the encrypted instruction C x Decrypt and restore the original planning instruction x * =D AES (C x ,K AES ) and transmitted to the exploration spacecraft.

[0036] The beneficial effects of the present invention are:

[0037] (1) Hybrid encryption mechanism: The method of using AES symmetric encryption combined with RSA asymmetric encryption mentioned in the present invention effectively protects the confidentiality and security of the image information obtained by the exploration spacecraft. This hybrid encryption mechanism improves the efficiency of data security transmission while taking into account the timeliness of encryption and decryption.

[0038] (2) Combining digital encryption technology with the Transformer perception model can effectively protect the security and privacy of aerospace data, enhance the robustness and reliability of the model, support multi-party secure computing and data collaboration, and thus improve the accuracy of state estimation and the overall performance of the system.

[0039] (3) Intelligent planning guidance combined with machine learning models: A machine learning model is used to plan the optimal observation position of the detection spacecraft, and combined with the optimization algorithm, a solution for automatically finding the optimal observation position is provided. The optimized optimal position is digitally encrypted to ensure its security during transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flowchart of a space target collaborative trusted perception method based on digital encryption provided by the present invention;

[0041] Figure 2 A schematic diagram of encryption of detection spacecraft images provided by the present invention;

[0042] Figure 3 A schematic diagram of decryption of a decision-making spacecraft image provided by the present invention;

[0043] Figure 4 Schematic diagram of the output result of the Transformer perception model provided by the present invention

[0044] Figure 5 A schematic diagram of the decision instruction encryption strategy provided by the present invention. DETAILED DESCRIPTION

[0045] Combination Figure 1-5 To illustrate this embodiment, Figure 1 As shown, the steps of a space target collaborative trusted perception method based on digital encryption described in this embodiment include:

[0046] S1: Encrypt the detection image information of the detection sensor carried by the detection spacecraft through digital encryption technology, and transmit the encrypted image and encryption public key to the decision-making spacecraft through an open channel;

[0047] like Figure 2 As shown, the encryption of detection spacecraft images includes the following steps:

[0048] S101: Use a randomly generated AES symmetric key K AES The image information I is encrypted, and the encrypted image information C is generated in the figure. I =E AES (I,K AES );

[0049] S102: Use the RSA public key of the receiving end For the AES symmetric key K AES Encrypt and generate a ciphertext symmetric key

[0050] S103: Transmit the encrypted image information C through a public channel I and the ciphertext symmetric key And use error correction code ECC and digital signature DS to improve the reliability and integrity of data transmission, that is, the transmission data is expressed as

[0051]

[0052] S104: To ensure the security of key storage and transmission, this embodiment periodically changes the AES symmetric key K AES and RSA key pair Select an appropriate key replacement cycle T, for example, every hour, every day, or every task cycle, to update the key. Through this series of steps, the security and privacy of image information can be effectively protected, and the reliability and integrity of data transmission can be guaranteed.

[0053] S2: The decision-making spacecraft uses the received encrypted image and encrypted public key as input, uses the variable L corresponding to the task as the supervision quantity to train the model, and outputs the variable required by the task

[0054] like Figure 3 As shown, decision-making spacecraft image decryption includes the following steps:

[0055] S201: Decision-making spacecraft uses its own RSA private key Symmetric key for ciphertext Decrypt and recover the AES symmetric key And use the recovered AES symmetric key K AES The encrypted image information C I Decrypt and restore the original image information I = D AES (C I ,K AES );

[0056] S202: Use the recovered AES symmetric key K AES The encrypted image information CI Decrypt and restore the original image information I = D AES (C I ,K AES );

[0057] S203: The decrypted image information I and variables L such as structural semantics and relative posture corresponding to the task are used as annotation data D=(I,L) to train the Transformer model;

[0058] During the training process, the image information I is used as the model input, and the task variable L is used as the supervision. The generalization ability of the model is improved through data enhancement techniques (e.g., rotation, scaling, cropping), and the Adam optimizer is used to minimize the loss function L and optimize the model parameters θ;

[0059] The update formula of Adam optimizer is:

[0060]

[0061] In formulas (1) and (2), β 1 is the decay rate of the first-order moment estimate, set to 0.9, is the gradient calculated at time step t, v t is the second-order moment estimate at time t, is the bias-corrected estimate of the second-order moment at time t, v t-1 is the second-order moment estimate at time t-1, β 2 is the decay rate of the second-order moment estimate, set to 0.999, m t is the first-order moment estimate at time t, is the bias-corrected estimate of the first-order moment at time t, m t-1 is the first-order moment estimate at time t-1, is the t-th power of the decay rate of the first-order moment estimate, is the t-th power of the decay rate of the second-order moment estimate, α is the initial learning rate, set to 0.001, and ε is a small constant to avoid dividing by 0 when updating parameters, set to 10 -8 ,θ t is the parameter vector at time t, θ t-1 is the parameter vector at time t-1.

[0062] S204: The trained Transformer model can output variables such as target structure semantics and relative pose required by the task like Figure 4 As shown, through this series of steps, it can be ensured that the decision-making spacecraft can accurately predict the variables required for the mission and improve the effectiveness and safety of the overall mission execution.

[0063] S3: Combining decision systems and variables Confirm the optimal observation position of the exploration spacecraft, encrypt the planning instructions and transmit them to the exploration spacecraft to achieve collaborative perception optimization of multiple spacecraft.

[0064] like Figure 5 As shown, the decision instruction encryption strategy includes the following steps:

[0065] S301: In order to find the optimal observation position of the detection spacecraft and ensure the security and privacy of the transmission instructions, we first define the implicit expression y=f(x) between the observation position x and the target observation result y, and use the reinforcement learning model to fit this relationship. The model output is By collecting training data sets for decision models where x i and i are the sequences representing different “observation positions-observation results” obtained through the test in step 2, and the loss function is used To train the model, use the Adam optimizer to update the parameters θ;

[0066] S302: Find the optimal observation position x by combining the optimization strategy * So that the objective function J(x) = -g(x,θ * ) is minimized, and the AES symmetric encryption algorithm is used to generate a random AES symmetric key K AES , for the optimal observation position x * Encrypt and generate encrypted instruction C x =E AES (x * ,K AES );

[0067] S303: Use RSA public key For the AES symmetric key K AES Encrypt and generate a ciphertext symmetric key Transmit the encrypted command C through an open channel x and the ciphertext symmetric key

[0068] S304: Decision-making spacecraft uses its own RSA private key Symmetric key for ciphertext Decrypt and recover the AES symmetric key And use the recovered AES symmetric key K AES For the encrypted instruction C x Decrypt and restore the original planning instruction x * =D AES (C x ,K AES) and transmit it to the exploration spacecraft. Through the above steps, it is ensured that the exploration spacecraft can obtain the optimal observation position, and the security and privacy of the instructions are guaranteed.

[0069] The above is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent replacement and improvement made to the above embodiments without departing from the content of the technical solution of the present invention, based on the technical essence of the present invention, within the spirit and principles of the present invention, still fall within the protection scope of the technical solution of the present invention.

Claims

1. A space target collaborative trusted perception method based on digital encryption, characterized in that: The steps of the space target collaborative trusted perception method based on digital encryption include: Step 1: Encrypt the detection image information of the detection sensor carried by the detection spacecraft through digital encryption technology, and transmit the encrypted image and encryption public key to the decision-making spacecraft through an open channel; Step 2: The decision-making spacecraft uses the received encrypted image and encrypted public key as input, uses the variable L corresponding to the task as the supervision quantity to train the model, and outputs the variable required by the task. Step 3: Combine decision systems and variables Confirm the optimal observation position of the exploration spacecraft, encrypt the planning instructions and transmit them to the exploration spacecraft to achieve collaborative perception optimization of multiple spacecraft.

2. According to the method of claim 1, the method is characterized in that: Step 1 specifically includes: Step 1.1: Use a randomly generated AES symmetric key K AES The detection data in the image information I is scrambled, forward diffused and reverse diffused to generate the encrypted image information C I =E AES (I,K AES ); Step 1.2: Use RSA public key For the AES symmetric key K AES Encrypt and generate a ciphertext symmetric key Step 1.3: Transmit the encrypted image information C through a public channel I To the decision-making spacecraft, and use the error correction code ECC and digital signature DS to the image information C I The transmission process is enhanced to generate transmission data Step 1.4: Pass the ciphertext symmetric key privacy transmission to decision-making spacecraft; Step 1.5: Regularly change the AES symmetric key K AES and RSA key pair in, is the RSA public key, To determine the spacecraft's own RSA private key, and to change the AES symmetric key K according to the set key replacement period T AES and RSA key pair to update.

3. According to the method of claim 1, the space target collaborative trusted perception method based on digital encryption is characterized in that: Step 2 specifically includes: Step 2.1: Use the decision-making spacecraft's own RSA private key Symmetric key for ciphertext Decrypt and recover the AES symmetric key Step 2.2: Use the recovered AES symmetric key For transmission data Decrypt and get the original image information I = D AES (C I ,K AES ) where the variable L includes structural semantics and relative posture; Step 2.3: Use the decrypted image information I and the variable L corresponding to the task as the labeled data D = (I, L) to train the Transformer perception model; Step 2.4: The trained Transformer perception model outputs the variables required by the task Among them, the variable Including the target structure semantics and relative posture required by the task.

4. According to the digital encryption-based space target collaborative trusted perception method of claim 3, it is characterized in that: Step 2.3 specifically includes: In the training process of the Transformer perception model, image information I is used as the model input and task variable L is used as the supervision variable. The generalization ability of the Transformer perception model is improved through data enhancement technology, and the Adam optimizer is used to minimize the loss function L to train the Transformer perception model.

5. According to the digital encryption-based space target collaborative trusted perception method of claim 1, it is characterized in that: Step 3 specifically includes: Step 3.1: Define the implicit expression y=f(x) between the observation position x of the detection spacecraft and the target observation result y, and use the reinforcement learning model to fit the implicit expression y to obtain the output result Step 3.2: Collect a training dataset for the decision model where x i and i are the sequences representing different "observation positions-observation results" obtained through the test in step 2, and the loss function is used To train the decision model based on reinforcement learning, the Adam optimizer is used to update the parameter θ; Step 3.3: Using optimization strategies and variables Get the optimal observation position x * Let the objective function J(x) = -g(x,θ * ) minimize; Step 3.4: Use the AES symmetric encryption algorithm to generate a random AES symmetric key K AES , for the optimal observation position x * Encrypt and generate encrypted instruction C x =E AES (x * ,K AES ); Step 3.5: Use RSA public key For the AES symmetric key K AES Encrypt and generate a ciphertext symmetric key Step 3.6: Transmit the encrypted command C through the public channel x and the ciphertext symmetric key Step 3.7: Decision spacecraft uses its own RSA private key Symmetric key for ciphertext Decrypt and recover the AES symmetric key And use the recovered AES symmetric key K AES For the encrypted instruction C x Decrypt and restore the original planning instruction x * =D AES (C x ,K AES ) and transmitted to the exploration spacecraft.

Images