A satellite telemetry sensing method based on fuzzy cognition

Abstract

This invention proposes a satellite telemetry sensing method based on fuzzy cognition. Through an event-driven organizational software architecture, the satellite's status is intelligently determined. When a corresponding situation is triggered on-board, the event triggers a sensing graph node and stores relevant fault telemetry data on-board. For complex-scale telemetry emergency faults in the corresponding constellation or on-board systems, accurate and rapid reasoning can be performed, generating corresponding decision information that is directly pushed to ground testing personnel and users, significantly improving fault location efficiency.

Description

Technical Field

[0001] This invention relates to a satellite telemetry sensing method based on fuzzy cognition, belonging to the field of satellite telemetry intelligent sensing technology. Background Technology

[0002] Satellite telemetry is fundamental to reflecting satellite health. With the explosive growth in the number of low-Earth orbit (LEO) constellations in the future, the types of satellite telemetry parameters are becoming increasingly complex, and their organizational structures are also more intricate. This results in a large amount of redundant and complex telemetry data being transmitted downlink, which not only inconveniences ground personnel in interpreting data and locating faults but also wastes onboard and ground memory resources and increases analytical pressure. LEO constellation networking is a future development trend, and the amount of complex telemetry information will increase accordingly, placing enormous pressure on ground testing personnel and users in determining satellite status and analyzing telemetry fault locations. An effective method is urgently needed to reduce the complexity of this problem. Fuzzy cognition is an intelligent computing method with numerical reasoning and formal representation characteristics. Its structure is similar to that of neural networks, combining fuzzy logic theory and neural networks for computation, and representing the logical relationships between concepts in the form of graphs.

[0003] The shortcomings of existing satellite telemetry sensing technologies are as follows: With constellation networking becoming the current trend, the volume of telemetry data from single and multiple satellites is increasing dramatically. Traditional overflight data transmission monitoring requires downloading telemetry data from all individual satellites, consuming significant resources. When all telemetry data from the on-orbit constellation is normal, a large amount of redundant and irrelevant data is downloaded. Furthermore, when a fault occurs on an on-orbit satellite, the faulty individual satellite cannot be quickly located, and delayed telemetry is required for auxiliary diagnosis, resulting in insufficient real-time performance and coordination in fault diagnosis. Summary of the Invention

[0004] The technical problem solved by this invention is to overcome the shortcomings of the prior art and provide a satellite telemetry sensing method based on fuzzy cognition. It uses a fuzzy cognitive learning algorithm to process on-board telemetry data. By interconnecting multi-satellite and single-satellite telemetry nodes and using an event-driven organizational software architecture, it intelligently determines the satellite status. When a corresponding situation is triggered on the satellite, the event triggers the sensing graph node and stores the relevant fault telemetry data on the satellite. It can accurately and quickly reason about complex-scale telemetry emergency faults in the corresponding constellation or on-board satellite and generate corresponding decision information, which is directly pushed to ground test personnel and users, which can greatly save telemetry transmission resources.

[0005] The technical solution of this invention is: a satellite telemetry sensing method based on fuzzy cognition, comprising: S1. Define the telemetry of the single-unit i-module on the satellite as a single-unit node. The set of nodes for each subsystem is denoted as . , And it is an integer, where n is the number of individual units of all subsystems on the satellite; Single-machine nodes in the model The state value at time t is denoted as... The larger the state value, the greater its impact on the entire system; construct a weight matrix representing the relationships between nodes on the satellite. This forms a fuzzy cognitive map; S2. When a single machine i fails, the single machine node is activated first. Then by a single node Trigger a single node single-machine node and single-machine node The connection is activated, and the state value of each node is updated based on the weight between nodes; After each iteration updates the state value, further adjustments are made to the single-machine node. and single-machine node Weight matrix between Update; S3, Determine the single-machine node If the state iteration standard value has reached the iteration learning stop standard, the iteration will return to S2 to continue iterating. If it has reached the standard, the iteration will terminate and an updated fuzzy cognitive map will be obtained. Based on this, decision-making information will be provided for the satellite and the ground. When a single unit on the satellite fails, the associated fault nodes will be quickly screened out and accurately located.

[0006] Preferably, the state values ​​of each individual node are:

[0007] in, For a single-machine node The state value at time t+1 after being affected by the node's own state value at time t and the state values ​​of other related nodes. It is an S-shaped threshold function. For a single-machine node The state value and weight matrix at time t. For a single node at time t For single-machine nodes The influence of weights.

[0008] Preferably, the weight matrix Indicates a single-machine node For single-machine nodes The influence of weights, , ;in: >0 represents a single node. For single-machine nodes There is a positive correlation. =0 represents a single node. For single-machine nodes No impact; <0 indicates a single node For single-machine nodes There is a negative correlation.

[0009] Preferably, for a single node and Weight matrix between The update will be performed as follows:

[0010] in, Scalar factor This represents a single node at time t+1. For single-machine nodes The influence of weights.

[0011] Preferably, a single-machine node State Iteration Standard F for:

[0012] in, To output the number of nodes, For a single-machine node The state value, For a single-machine node The midpoint of the target state interval, i.e.:

[0013] in, It is a single-machine node The target state interval, For a single-machine node The minimum value of the target state interval, For a single-machine node The maximum value of the target state interval; The stopping criterion for iterative learning is: when The iteration terminates when the value is less than 0.001.

[0014] Preferably, at the initial moment, each individual node in the single-star model... Status values ​​and single-machine nodes right weights Based on prior information or experience, it is determined to be an unsupervised fuzzy cognitive learning algorithm based on Hebbian.

[0015] Secondly, a terminal device is provided, comprising: Memory, used to store at least one instruction executed by a processor; A processor is used to execute instructions stored in memory to implement the methods described above.

[0016] Thirdly, a computer-readable storage medium is provided that stores computer instructions that, when executed on a computer, cause the computer to perform the methods described above.

[0017] Compared with the prior art, the present invention has the following advantages: (1) By means of fuzzy cognitive learning algorithms, it is possible to quickly locate the faults triggered by on-board telemetry, and store the most relevant data information for ground telemetry, thus eliminating a large amount of redundant and complex information for users; (2) It also alleviates the pressure on satellite resources and ground analysis and on-orbit management, and optimizes resource storage. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method of the present invention; Figure 2 The fuzzy cognitive map provided in the embodiments of the present invention. Detailed Implementation

[0019] This invention addresses the challenges of telemetry display for future satellite constellations, rapid fault location, and improved user application efficiency. It proposes a satellite telemetry sensing method based on fuzzy cognition. In normal on-orbit mode, it presents only key telemetry data of the constellation, such as time, orbit, and attitude, to the user. When a fault occurs on the satellite, it intelligently fuses complex on-board telemetry data based on event-driven processing. By activating corresponding on-board nodes through a fuzzy cognitive map, it quickly locates relevant telemetry data and directly transmits the most relevant telemetry data to the ground, eliminating other redundant and irrelevant data, thus significantly improving fault location efficiency.

[0020] This invention is based on the ability of fuzzy cognition to simulate the dynamic operation of satellite telemetry. The constellation and on-board telemetry sensing process is the process of iterative calculation of the weight matrix in the fuzzy cognitive graph.

[0021] Initial satellite telemetry values ​​can be obtained by abstracting expert knowledge, extracting telemetry data from different individual units and systems into concept nodes in a fuzzy cognitive graph, and assigning weights to the influence between concept nodes based on experience, thus obtaining a preliminary fuzzy cognitive graph library.

[0022] The accuracy of fuzzy cognitive graph reasoning is improved by learning and optimizing the fuzzy cognitive graph library using a nonlinear Hebbian algorithm. The platform is comprehensively validated in a software environment, and the fuzzy cognitive graph library is continuously optimized through self-learning using the nonlinear Hebbian algorithm, progressively improving the accuracy of decision-making results.

[0023] Specifically, a satellite telemetry sensing method based on fuzzy cognition includes: The telemetry of the individual i-modules of each subsystem on the satellite is defined as a node. The set of nodes is denoted as , And it is an integer, where n is the number of individual units of all subsystems on the satellite; Single-machine nodes in the model The state value at time t is denoted as... The larger the state value, the greater its impact on the entire system; based on this, a fuzzy cognitive graph is constructed to establish the connection relationships between nodes on the satellite, and the set of weight matrices is as follows:

[0024] Among them, the weight matrix Indicates a single-machine i-node For a single j-node The influence of weights, , ;in, >0 represents standalone operation. For standalone machines There is a positive correlation. =0 represents standalone operation. For standalone machines No impact; <0 indicates standalone operation For standalone machines There is a negative correlation effect; When a single node i on the satellite malfunctions, the single node i node is activated first. Then by a single node Trigger single-player Only a single node and single-machine node The connection is activated, and the other unrelated nodes do not need to be activated. The state value of each node is updated based on the weight between the nodes. After updating the state values ​​in each iteration, the weight matrix between the connected nodes is further adjusted. Update; Determine the status value of a single node If the state iteration standard value has reached the iteration learning stopping standard, the iteration will continue if it has not reached the standard. If it has reached the standard, the iteration will terminate and an updated fuzzy cognitive map will be obtained. Based on this, decision-making information can be provided for the satellite and the ground. When a single unit on the satellite fails, the associated fault nodes can be quickly screened out and accurately located.

[0025] (1) The status values ​​of each individual node are:

[0026] in, For a single-machine node The state value at time t+1 after being affected by the node's own state value at time t and the state values ​​of other related nodes. It is an S-shaped threshold function. For a single-machine node The state value at time t, For a single node at time t For single-machine nodes The influence of weights.

[0027] (2) Weight matrix between connected nodes The update will be performed as follows:

[0028] in, Scalar factor This represents a single node j at time t+1. For a single-machine i-node The influence of weights.

[0029] (3) Determine the status value of a single node When the state iteration standard value reaches the iterative learning stopping criterion:

[0030] in, To output the number of nodes, For a single-machine node The state value, For a single-machine node The midpoint of the target state interval, i.e.:

[0031] in, It is a single-machine node The target state interval. When The iteration terminates when the minimum value is reached, i.e., F < 0.001.

[0032] Otherwise, continue iterating over the weight matrix to perform iterative learning.

[0033] (4) At the initial moment, each single-machine node in the single-star model State values ​​and nodes right weights Based on prior information or experience, it is an unsupervised Hebbian-based fuzzy cognitive learning algorithm.

[0034] Example: This invention provides a satellite telemetry sensing method based on fuzzy cognition, such as... Figure 1 and Figure 2 As shown, it includes the following steps: 1) Individual nodes of each subsystem for single-satellite telemetry data ,in Indicates the first A single-unit conceptual node. Taking the control subsystem of single satellite A as an example, it includes key telemetry parameters and fault modes such as propulsion subsystem module C1, fiber optic gyroscope C2, star sensor C3, control torque gyroscope C4, magnetic moment sensor C5, digital sun sensor C6, control system-level telemetry C7, and control torque gyroscope break-in processing C8. The telemetry of the control subsystems of the main satellite, single satellite A, single satellite B, and single satellite C in the constellation is similar.

[0035] 2) The telemetry of the single-satellite control subsystem is connected by directed arcs to link the key telemetry parameter nodes of each component. Fuzzy reasoning is represented by the relationship between concept nodes. The fuzzy cognitive graph is a network with direction and symbols, which is composed of directed connection arcs between multiple nodes such as C1, C2, C3, C4, C5, C6, C7, and C8.

[0036] 3) Adjacency connection matrix of fuzzy cognitive telemetry nodes in a single-satellite control subsystem .

[0037] in, For a single-machine node right The influence of weights.

[0038] 4) Similarly, the fuzzy cognition of a multi-satellite telemetry node involves treating each individual unit as a whole as an important node, and then interconnecting with the main satellite through these nodes. The adjacency connection matrix of a multi-satellite fuzzy cognition telemetry node. .

[0039] 5) The state value of each node in the model at time t is The larger the state value, the greater the impact on the system. Based on a fuzzy cognitive learning method using satellite telemetry, when an on-orbit single-unit control subsystem enters a fault mode, it is detected through event nodes. Trigger And record time information and related telemetry data, then the node and connection matrix Once activated, the remaining unrelated nodes do not need to be activated.

[0040] 6) The satellite telemetry sensing method uses a nonlinear Hebbian learning algorithm. This algorithm continuously optimizes the weights between convergent nodes, resulting in the following fuzzy cognitive state values ​​for each node:

[0041] in, For nodes The state value at time t+1 after being affected by the node's own state value at time t and the state values ​​of other related nodes. This is an S-shaped threshold function. The output nodes represent the final states of each module in the satellite telemetry system.

[0042] 7) After each iteration of the state value, the weights of the arcs connecting all nodes must be updated:

[0043] in, It is a scalar factor.

[0044] 8) The stopping criterion for iterative learning is:

[0045] in, To output the number of nodes, For a single-machine node The state value, For a single-machine node The midpoint of the target state interval, i.e.:

[0046] in, It is a single-machine node The target state interval. When The iteration terminates when the value is less than 0.001.

[0047] The algorithm determines whether the nonlinear Hebbian learning algorithm has reached the stopping criterion for iterative learning. If not, it returns to step 7) to continue iterating the weight matrix and updating the state values. If the stopping criterion is met, the iteration terminates, resulting in an updated fuzzy cognitive map, which provides decision-making information for both spacecraft and ground systems. This method is suitable for iterative learning problems of fuzzy cognitive maps supported by expert experience.

[0048] This invention adopts Figure 1 The illustrated process completes a satellite telemetry sensing method based on fuzzy cognition. Through a nonlinear Hebbian learning algorithm, the weights between convergent nodes are continuously optimized to obtain the final state vector, which then pushes key telemetry information to ground testing personnel and users. The invention will be described in detail below with specific examples, such as... Figure 2 As shown: Decompose each unit into nodes Each node has a corresponding weight pointing to a matrix. This value reflects the strength of the causal relationship between nodes. The control subsystem is connected through individual nodes, and fuzzy inference is demonstrated through conceptual nodes. For example, if a high-speed or low-speed fault occurs in the on-orbit CMG, node C4 is directly triggered, and corresponding nodes C7 and C8 are triggered through weight allocation. C7 is for system telemetry, and C8 is for fault handling CMG break-in. Other unrelated nodes are not activated, and there is no need for telemetry or user push notifications, which can greatly save on-board resources.

[0049] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solutions of the present invention by utilizing the methods and techniques disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention shall fall within the protection scope of the technical solutions of the present invention.

[0050] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A satellite telemetry sensing method based on fuzzy cognition, characterized in that... include: S1. Define the telemetry of the single-unit i-module on the satellite as a single-unit node. The set of nodes for each subsystem is denoted as . , And it is an integer, where n is the number of individual units of all subsystems on the satellite; Single-machine nodes in the model The state value at time t is denoted as... The larger the state value, the greater its impact on the entire system; construct a weight matrix representing the relationships between nodes on the satellite. This forms a fuzzy cognitive map; S2. When a single machine i fails, the single machine node is activated first. Then by a single node Trigger a single node single-machine node and single-machine node The connection is activated, and the state value of each node is updated based on the weight between nodes; After each iteration updates the state value, further adjustments are made to the single-machine node. and single-machine node Weight matrix between Update; S3, Determine the single-machine node If the state iteration standard value has reached the iteration learning stop standard, the iteration will return to S2 to continue iterating. If it has reached the standard, the iteration will terminate and an updated fuzzy cognitive map will be obtained. Based on this, decision information will be provided for the satellite and the ground. When a single unit on the satellite fails, the associated fault nodes will be quickly screened out and accurately located.

2. The satellite telemetry sensing method based on fuzzy cognition according to claim 1, characterized in that: The status values ​​of each individual node are: in, For a single node The state value at time t+1 after being affected by the node's own state value at time t and the state values ​​of other related nodes. It is an S-shaped threshold function. For a single node The state value and weight matrix at time t. For a single node at time t For single-machine nodes The influence of weights.

3. The satellite telemetry sensing method based on fuzzy cognition according to claim 1, characterized in that: weight matrix Indicates a single-machine node For single-machine nodes The influence of weights, , ;in: >0 represents a single node. For single-machine nodes There is a positive correlation. =0 represents a single node. For single-machine nodes No impact; <0 indicates a single node For single-machine nodes There is a negative correlation.

4. The satellite telemetry sensing method based on fuzzy cognition according to claim 1, characterized in that: For single-machine nodes and Weight matrix between The update will be performed as follows: in, Scalar factor This represents a single node at time t+1. For single-machine nodes The influence of weights.

5. The satellite telemetry sensing method based on fuzzy cognition according to claim 1, characterized in that: single-machine node State Iteration Standard F for: in, To output the number of nodes, For a single node The state value, For a single node The midpoint of the target state interval, i.e.: in, It is a single-machine node The target state interval, For a single node The minimum value of the target state interval, For a single node The maximum value of the target state interval; The stopping criterion for iterative learning is: when The iteration terminates when the value is less than 0.

001.

6. The satellite telemetry sensing method based on fuzzy cognition according to claim 1, characterized in that: At the initial moment, each individual node in the single-star model Status values ​​and single-machine nodes right weights Based on prior information or experience, it is determined to be an unsupervised fuzzy cognitive learning algorithm based on Hebbian.

7. A terminal device, characterized in that, include: Memory, used to store at least one instruction executed by a processor; A processor for executing instructions stored in memory to implement the method as described in any one of claims 1-6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1-6.

Images