A method, apparatus, and electronic equipment for evaluating the performance of a satellite system.

Abstract

This application discloses a method, apparatus, and electronic device for evaluating the performance of a satellite system. The method includes: acquiring performance parameter values ​​of the satellite system to be evaluated on different specified performance indicators; acquiring the target value range of each specified performance indicator according to a pre-constructed mapping relationship; the mapping relationship being the correspondence between performance indicators and value ranges; the value range being the actual range of values ​​that the corresponding performance indicator can take during normal operation of the satellite system; and determining the dimension score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance indicator based on the performance parameter values ​​and the target value range, thereby determining whether the performance of the satellite system to be evaluated meets expectations. Thus, by comparing the actually acquired performance parameter values ​​with the preset actual range of values, it can be determined whether the actual performance parameter values ​​are within the acceptable range, thereby enabling the scoring and evaluation of the satellite system on the evaluation dimensions corresponding to different specified performance indicators to analyze the satellite system performance.

Description

Technical Field

[0001] This application relates to the field of satellite system technology, and in particular to a method, apparatus and electronic equipment for evaluating the performance of a satellite system. Background Technology

[0002] As satellite constellations expand, traditional ground-based central control-based satellite management models are facing bottlenecks. Specifically, at the constellation autonomy level, energy scheduling strategies are mostly statically preset, resulting in limited energy utilization; thermal control systems experience temperature fluctuations under extreme conditions; and fault repair heavily relies on ground intervention, leading to long recovery cycles.

[0003] Furthermore, at the constellation coordination level, mission scheduling relies heavily on pre-programmed rules, and the injection of emergency missions and replanning are time-consuming, with weak dynamic revisit optimization capabilities, resulting in limited overall coordination efficiency and fault tolerance of the satellite system. At the on-orbit computing level, onboard models suffer from high inference latency, insufficient incremental learning capabilities, and poor adaptability to identifying unknown targets due to resource constraints.

[0004] Currently, although local optimization can be achieved through methods such as model lightweighting or dedicated hardware acceleration, the lack of effective verification of optimization techniques due to the inability to evaluate the performance of satellite system autonomy, collaboration, and on-orbit computing has hindered the overall improvement of satellite system intelligence.

[0005] Therefore, how to evaluate the performance of satellite systems in terms of autonomous management, collaborative control, and on-orbit computing, and improve the overall efficiency and reliability of space mission execution, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] In view of this, one aspect of this application provides a method for evaluating the performance of a satellite system, the method comprising: Obtain the performance parameter values ​​of the satellite system to be evaluated on a specified performance index; the specified performance index includes at least one of autonomous index, control index, and on-orbit computation index. Based on the pre-constructed mapping relationship, the target value range of each specified performance indicator is obtained; the mapping relationship is the correspondence between the performance indicator and the value range; the value range is the actual range of values ​​that the corresponding performance indicator can take when the satellite system is operating normally. Based on the performance parameter values ​​and the target value range, determine the dimension score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index; Based on the scores of the aforementioned dimensions, it is determined whether the performance of the satellite system to be evaluated meets expectations.

[0007] Optionally, determining the dimensional score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index based on the performance parameter value and the target value range includes: Obtain a pre-built evaluation model; the evaluation model includes a first model and a second model; the first model is a piecewise function that monotonically increases with respect to the performance index; the second model is a piecewise function that monotonically decreases with respect to the performance index. Determine whether the specified performance indicators are positively correlated with the satellite system performance; If so, the dimension score is determined using the first model based on the performance parameter value and the target value range; If not, the dimension score is determined using the second model based on the performance parameter values ​​and the target value range.

[0008] Optionally, determining the dimension score based on the performance parameter value and the target value range using the first model includes: If the performance parameter value is greater than the maximum value in the target value range, the first score will be used as the dimension score. If the performance parameter value is within the target value range, a second score is determined based on the maximum value, the minimum value within the target value range, and the performance parameter value; and the second score is used as the dimension score. If the performance parameter value is less than the minimum value, the third score is used as the dimension score; wherein the first score is greater than the second score; and the second score is greater than the third score.

[0009] Optionally, determining the dimension score using the second model based on the performance parameter value and the target value range includes: If the performance parameter value is greater than the maximum value, the third score will be used as the dimension score. If the performance parameter value is within the target range, the second score will be used as the dimension score. If the performance parameter value is less than the minimum value, the first score is used as the dimension score.

[0010] Optionally, determining the second score based on the maximum value, the minimum value in the target value range, and the performance parameter value includes: Obtain the boundary values ​​of the pre-set evaluation model output results; the boundary values ​​include a maximum boundary value and a minimum boundary value, wherein the maximum boundary value is less than the first score; and the minimum boundary value is greater than the third score; The key parameter values ​​of the evaluation model are determined by the boundary values ​​and the target value range. The second score is determined based on the key parameter values ​​and the performance parameter values.

[0011] Optionally, the key parameter values ​​include a first parameter value and a second parameter value; The first parameter value is determined based on the boundary value, the maximum value, and the minimum value. In the first model, the first parameter value is less than 0, and in the second model, the first parameter value is greater than 0. In the first model, the second parameter value is determined based on the minimum value, the first parameter value, and the minimum boundary value; In the second model, the second parameter value is determined based on the maximum value, the first parameter value, and the minimum boundary value.

[0012] Optionally, the autonomous indicators include energy utilization rate, temperature stability index, autonomous repair rate, system fault recovery time, peak model memory usage, and fault diagnosis power consumption; The control indicators include mission planning duration, revisit cycle reduction rate, energy utilization rate improvement rate, anomaly handling success rate, and satellite positioning error. The on-orbit computing metrics include target detection accuracy, model inference latency, incremental learning decay, model energy efficiency ratio, and unknown target recall rate.

[0013] Optionally, the specified performance indicators include the autonomous indicator, the control indicator, and the on-orbit computation indicator; determining whether the performance of the satellite system to be evaluated meets expectations based on the dimensional scores includes: First weights are assigned to the first dimension scores corresponding to the autonomous indicators; second weights are assigned to the second dimension scores corresponding to the control indicators; and third weights are assigned to the third dimension scores corresponding to the on-orbit computing indicators; the sum of each first weight, the sum of each second weight, and the sum of each third weight are all 1. Based on the first weight, the scores of the first dimension are weighted and summed to obtain the autonomous layer score; Based on the second weight, the scores of the second dimension are weighted and summed to obtain the control layer score; Based on the third weight, the scores of the third dimension are weighted and summed to obtain the on-orbit computing layer score; Based on the scores of the autonomous layer, the control layer, and the orbit calculation layer, it is determined whether the performance of the satellite system to be evaluated meets expectations.

[0014] Optionally, determining whether the performance of the satellite system to be evaluated meets expectations based on the autonomous layer score, the control layer score, and the orbit calculation layer score includes: A first coefficient is configured for the score of the autonomous layer, a second coefficient is configured for the score of the control layer, and a third coefficient is configured for the score of the orbit calculation layer; the sum of the first coefficient, the second coefficient, and the third coefficient is 1. Based on the first coefficient, the second coefficient, and the third coefficient, the scores of the autonomous layer, the control layer, and the orbit calculation layer are weighted and summed to obtain a comprehensive score. When the overall score is greater than the threshold, it is determined that the performance of the satellite system to be evaluated meets expectations.

[0015] Another aspect of this application provides a satellite system performance evaluation device, the device comprising: The parameter value acquisition module is used to acquire the performance parameter values ​​of the satellite system to be evaluated on a specified performance index; the specified performance index includes at least one of autonomous index, control index and on-orbit calculation index. The value range determination module is used to obtain the target value range of each specified performance indicator according to a pre-constructed mapping relationship; the mapping relationship is the correspondence between the performance indicator and the value range; the value range is the actual range of values ​​that the corresponding performance indicator can take when the satellite system is operating normally. The dimension score determination module is used to determine the dimension score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index based on the performance parameter value and the target value range. The performance evaluation module is used to determine whether the performance of the satellite system to be evaluated meets expectations based on the scores of the dimensions.

[0016] Another aspect of this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor executes the computer program to implement the steps of a method for evaluating the performance of the satellite system.

[0017] The satellite system performance evaluation method, apparatus, and electronic equipment provided in this application have the following beneficial effects: After obtaining the actual acquired performance parameter values ​​and the pre-set range of acceptable values, the comparison between the two can determine whether the actual performance parameter values ​​of the satellite system are within the acceptable range, thereby scoring and evaluating the satellite system on the evaluation dimensions corresponding to different specified performance indicators. A standardized and quantifiable performance evaluation system is established, enabling performance evaluation of the satellite system on different dimensions such as autonomous indicators, control indicators, and on-orbit computing indicators. This system can be used to verify the reliability and real-time performance of the satellite system under different operating environments, improving the overall efficiency and reliability of space mission execution. Attached Figure Description

[0018] Figure 1A flowchart illustrating a method for evaluating the performance of a satellite system provided in an embodiment of this application; Figure 2 A schematic diagram of the structure of a satellite system performance evaluation system provided in an embodiment of this application; Figure 3 A flowchart illustrating a method for evaluating the performance of a satellite system, provided as another embodiment of this application; Figure 4 A schematic diagram of the structure of a satellite system performance evaluation device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0019] The attached diagram is labeled as follows: 40 is the parameter value acquisition module, 41 is the value range determination module, 42 is the dimension score determination module, 43 is the performance evaluation module, 50 is the memory, 51 is the processor, 52 is the display screen, 53 is the input / output interface, 54 is the communication interface, 55 is the power supply, 56 is the communication bus, 501 is the computer program, 502 is the operating system, and 503 is the data. Detailed Implementation

[0020] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0021] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0022] Figure 1 This is a flowchart illustrating a method for evaluating the performance of a satellite system provided in an embodiment of this application, as shown below. Figure 1 As shown, the method includes: S10: Obtain the performance parameter values ​​of the satellite system to be evaluated on the specified performance indicators; the specified performance indicators include at least one of the following: autonomous indicators, control indicators, and on-orbit calculation indicators; Figure 2This is a schematic diagram of a satellite system performance evaluation system provided in an embodiment of this application. It is understood that, in specific embodiments, the ground-centric control-based satellite management model has problems at three levels: constellation autonomy, constellation coordinated control, and on-orbit computing. To achieve performance evaluation of the satellite system at different levels, such as... Figure 2 As shown, in one alternative embodiment, the evaluation system can evaluate the satellite system to be evaluated along three paths.

[0023] For details, see Figure 2 The three paths are Path 1, Path 2, and Path 3. Path 1 is used to evaluate the performance of the satellite system under evaluation at the constellation autonomy level, Path 2 is used to evaluate the performance of the satellite system under evaluation at the constellation coordinated control level, and Path 3 is used to evaluate the performance of the satellite system under evaluation at the on-orbit computing level.

[0024] In a specific embodiment of the performance evaluation of the satellite system to be evaluated, the actual performance parameter values ​​of the satellite system to be evaluated on specified performance indicators are obtained. The satellite system to be evaluated may be, but is not limited to, a low-Earth orbit remote sensing constellation, a communication constellation, a navigation constellation, and a computing constellation. The actual performance parameter values ​​collected may be performance parameter values ​​generated in real time by the operating satellite system, or performance parameter values ​​collected when controlling the satellite system to be evaluated to operate under specified parameter environments; this application does not limit the scope of these values.

[0025] It is worth noting that the specified performance indicators can include at least one of the following: autonomous indicators, control indicators, and on-orbit computation indicators. Specifically, the selection can be made according to actual operational needs. When multiple specified performance indicators are selected, a comprehensive evaluation of the performance of the satellite system under evaluation across multiple dimensions can be conducted.

[0026] It should be noted that autonomous indicators refer to indicators related to the autonomous operation of the satellite system, including but not limited to energy utilization rate, temperature stability, autonomous repair rate, and system fault recovery time. Control indicators refer to indicators related to the coordinated control between satellite systems, including but not limited to mission planning time, revisit cycle reduction rate, and energy utilization rate improvement rate. On-orbit computing indicators refer to indicators related to the real-time computing of the satellite system, including but not limited to target detection accuracy, model inference latency rate, and incremental learning decay.

[0027] Correspondingly, based on the above indicators, different data are collected at different layers to obtain the corresponding performance parameter values. Table 1 is a schematic table illustrating the correspondence between evaluation dimensions and performance parameter values ​​provided in the embodiments of this application. For ease of understanding, the following explanation will be based on Table 1.

[0028] Table 1 shows a schematic diagram of the correspondence between one evaluation dimension and performance parameter values.

[0029] See Table 1 and Figure 2 As can be seen, in specific embodiments, multiple rating dimensions are defined for the constellation autonomous layer, the collaborative control layer, and the on-orbit computing layer to evaluate the performance of the satellite system. Different evaluation dimensions correspond to different specified performance indicators. For example, for the energy utilization rate evaluation dimension, the corresponding specified performance indicator is energy utilization rate, and the performance parameter value is the specific numerical value of energy utilization rate.

[0030] In a specific embodiment, real-time data of the satellite system to be evaluated is collected based on the description on the far right, and performance parameter values ​​are calculated using the real-time data. In an optional embodiment, the constellation autonomous layer collects energy data (e.g., solar panel output current / voltage, battery charge / discharge curves, etc.), thermal control data (e.g., internal / external temperature gradient, heat sink efficiency, etc.), and fault logs (e.g., sensor failure codes, communication interruption sequences, etc.) in real time through onboard sensors.

[0031] The collaborative control layer collects task scheduling timing (such as orbital maneuver commands, payload on / off status, etc.), multi-satellite collaborative data (such as staring positioning coordinate differences, resource allocation records, etc.), and anomaly injection responses (such as attitude loss correction trajectory, data retransmission success rate, etc.), energy data, etc.

[0032] The on-orbit computing layer focuses on the image processing pipeline, acquiring raw remote sensing images, model inference latency logs, incremental learning parameter updates, and unknown target recognition results. Specific acquisition frequencies and calculation methods are detailed in Table 1.

[0033] S11: Based on the pre-built mapping relationship, obtain the target value range of each specified performance indicator; the mapping relationship is the correspondence between the performance indicator and the value range; the value range is the actual range of values ​​that the corresponding performance indicator can take when the satellite system is operating normally. Understandably, during normal operation of a satellite system, different specified performance indicators can take on values ​​within a realistic and reasonable range. In other words, for different specified performance indicators, if the satellite system is operating normally, there will be a corresponding reasonable range of values ​​that conforms to reality. In fact, this range can also be understood as the range of values ​​that can be used to evaluate the performance of the satellite system being evaluated. For ease of understanding, examples will be provided below.

[0034] For example, regarding the energy utilization rate (EUE) of a satellite system, in a specific embodiment, if the EUE is less than 10%, the satellite system has very serious performance problems in the actual operation scenario, and the evaluation of the satellite system is meaningless at this time.

[0035] If the EUE value is between 10% and 30%, the satellite system is operating normally in actual satellite operation scenarios, meaning the satellite system performance is normal. However, different values ​​correspond to different performance levels. Furthermore, the satellite system performs optimally when the EUE value is 25%. Therefore, the actual and reasonable range for EUE can be pre-set to 10%-30%.

[0036] If the EUE is greater than 30%, it is impossible to reach this value range in the actual operation scenario of the satellite. If this value range is still used for performance evaluation, it may result in a satellite system with good actual performance but give a bad evaluation result, which is seriously deviated from the real-time situation and does not conform to the real scenario, thus affecting the final score.

[0037] Based on the above analysis, in one optional embodiment, a mapping relationship is obtained by pre-constructing the correspondence between performance indicators and value ranges. The mapping relationship can be stored in the form of a table or in other forms, and this application does not limit this.

[0038] Therefore, based on the performance parameter values ​​obtained above, and further, based on the mapping relationship, the target value range corresponding to each specified performance indicator is determined. For example, referring to the example above, if the specified performance indicator includes EUE, the corresponding target value range is 10%-30%.

[0039] S12: Based on the performance parameter values ​​and the target value range, determine the dimension score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index; By obtaining the actual parameters (i.e., performance parameter values) and the reasonable value range (i.e. target value range) through steps S10 and S11, the dimensional score of the satellite system to be evaluated on the corresponding evaluation dimension can be determined by comparing the two.

[0040] In one optional embodiment, the performance parameter values ​​and the target value range can be input into a designated large model, which then scores the model on the corresponding evaluation dimension to obtain a dimension score. The designated large model may include, but is not limited to, the GPT (Generative Pre-trained Transformer) series, the BERT model, large models based on the Transformer architecture, LLaMA (Large Language Model Meta AI), and Tongyi Qianwen. This application does not limit the target large model.

[0041] In one optional embodiment, the performance parameter value is compared with a target value range, and different dimension scores are pre-set for different comparison results. Specifically, another mapping relationship is pre-constructed, which includes the correspondence between the comparison results and the dimension scores. The comparison result refers to the comparison between the performance parameter value and the target value range, specifically including performance parameter values ​​within the target value range, performance parameter values ​​greater than the maximum value in the target value range, and performance parameter values ​​less than the minimum value in the target value range.

[0042] In one optional embodiment, a first score is obtained when the performance parameter value is greater than the maximum value in the target range, and a second score is obtained when the performance parameter value is less than the minimum value in the target range. When the performance parameter value is within the target range, the target range can be divided into multiple value segments, with different value segments corresponding to different scores.

[0043] In another optional embodiment, a fuzzy evaluation model can be pre-constructed, and the dimensional score can be directly calculated using this fuzzy evaluation model based on the performance parameter values ​​and the target value range. This application does not limit the construction of the fuzzy evaluation model.

[0044] S13: Based on the dimensional scores, determine whether the performance of the satellite system to be evaluated meets expectations.

[0045] After obtaining the dimensional scores for each of the multiple evaluation dimensions, the dimensional scores are summed to obtain a comprehensive score for the satellite system under evaluation. See also: Figure 2 The system calculates scores for different layers for different paths. Specifically, the constellation autonomous layer obtains an autonomous layer score, the collaborative control layer obtains a control layer score, and the on-orbit computing layer obtains an on-orbit computing layer score. Furthermore, the performance of different layers can be evaluated separately, or the scores of the three layers can be further evaluated to obtain a comprehensive score for analyzing the overall performance of the system. This application does not limit the scope of this application.

[0046] Therefore, the satellite system performance evaluation method provided in this application, after obtaining the actual acquired performance parameter values ​​and the pre-set actual acceptable value range, can determine whether the actual performance parameter values ​​of the satellite system are within the acceptable value range based on the comparison results of the two, thereby scoring and evaluating the satellite system on the evaluation dimensions corresponding to different specified performance indicators. A standardized and quantifiable performance evaluation system is established, enabling performance evaluation of the satellite system on different dimensions such as autonomous indicators, control indicators, and on-orbit computing indicators. This can be used to verify the reliability and real-time performance of the satellite system under different operating environments, improving the overall efficiency and reliability of space mission execution.

[0047] Figure 3 This is a flowchart illustrating a satellite system performance evaluation method according to another embodiment of this application. In an optional embodiment, such as... Figure 3 As shown, based on the performance parameter values ​​and the target value range, the dimensional scores of the satellite system to be evaluated on the evaluation dimensions corresponding to the specified performance indicators are determined, including: S30: Obtain the pre-built evaluation model; the evaluation model includes a first model and a second model; the first model is a piecewise function with respect to a monotonically increasing performance index; the second model is a piecewise function with respect to a monotonically decreasing performance index. S31: Determine whether the specified performance index is positively correlated with the satellite system performance; if yes, proceed to step S32; if no, proceed to step S33. S32: Based on the performance parameter values ​​and the target value range, determine the dimension score using the first model; S33: Based on the performance parameter values ​​and the target value range, determine the dimension score through the second model.

[0048] In a specific embodiment, a fuzzy evaluation model is pre-constructed, and the performance parameter values ​​acquired in real time are input into the evaluation model to calculate the dimensional score. Specifically, in an optional embodiment, the evaluation model includes a first model and a second model.

[0049] The output of the first model monotonically increases with the performance index, meaning that the larger the performance index, the larger the output of the first model, and the better the performance of the corresponding satellite system. The output of the second model monotonically decreases with the performance index, meaning that the smaller the performance index, the larger the output of the second model, and the better the performance of the corresponding satellite system.

[0050] Based on this, in a specific embodiment, it is determined whether the specified performance index is positively correlated with the satellite system performance, that is, whether the larger the specified performance index is, the better the satellite system performance is, so that the corresponding evaluation model can be selected to calculate the dimensional score.

[0051] Specifically, if a larger specified performance index corresponds to better system performance (i.e., a positive correlation), the first model, which is monotonically increasing, is selected for calculation. If a smaller specified performance index corresponds to better system performance (i.e., a negative correlation), the second model, which is monotonically increasing, is selected for calculation.

[0052] Based on the above embodiments, as an optional embodiment, the dimension score is determined through the first model according to the performance parameter values ​​and the target value range, including: If the performance parameter value is greater than the maximum value in the target range, the first score will be used as the dimension score. If the performance parameter value is within the target range, the second score is determined based on the maximum value, the minimum value within the target range, and the performance parameter value; and the second score is used as the dimension score. If the performance parameter value is less than the minimum value, the third score will be used as the dimension score; where the first score is greater than the second score, and the second score is greater than the third score.

[0053] In a specific embodiment, the first model monotonically increases with the performance index, meaning that the larger the collected performance parameter value, the larger the output of the first model. In an optional embodiment, the first model can be constructed as formula (1): (1) in, For the first model, For first place, For second place, Third place score, These are performance parameter values. The maximum value within the target's range. The minimum value within the target range.

[0054] It should be noted that, in one optional embodiment, the first score... and third score The second score is a pre-set specific value. Further analysis based on the maximum value is needed. Minimum value and performance parameter values Real-time calculations are performed. The first score is... Greater than the second score Second score Greater than the third score .

[0055] As can be seen from formula (1), in a specific embodiment, if the performance parameter value Greater than the maximum value Since the first model is about performance parameter values The function is monotonically increasing, at which point the model's output is maximized. In one optional embodiment, the first score... It can take the value 1.

[0056] If the performance parameter value Less than the minimum value Since the first model is about performance parameter values The function is monotonically increasing, at which point the model's output is minimized. In one optional embodiment, the third score... It can take the value 0. That is, the first model. The output value ranges from 0 to 1.

[0057] If the performance parameter value If the value is within the target range, further calculation is required, and the calculation result should still conform to the performance parameter value. The rule is that the larger the value, the larger the output result.

[0058] In one optional embodiment, based on the performance parameter values ​​and the target value range, a second model is used to determine the dimension score, including: If the performance parameter value is greater than the maximum value, the third score will be used as the dimension score; If the performance parameter value is within the target range, the second score will be used as the dimension score. If the performance parameter value is less than the minimum value, the first score will be used as the dimension score.

[0059] In a specific embodiment, the second model monotonically decreases with the performance index, meaning that the smaller the collected performance parameter value, the larger the output of the first model. In an optional embodiment, the second model can be constructed as formula (2): (2) in, This is the first model.

[0060] Based on the analysis of the above calculation formula (1), see formula (2), since the first model With performance parameter values Monotonically decreasing. Therefore, when the performance parameter value... The larger the value, the smaller the output, meaning it's closer to the third-place score. (For example, the closer to 0). Conversely, when the performance parameter value... The smaller the value, the larger the output, meaning it's closer to the first score. (For example, the closer to 1).

[0061] Similarly, performance parameter values When the target value is within the range, it is also necessary to consider the maximum value. Minimum value and performance parameter values Real-time calculations are performed, and the results at this point match the performance parameter values. The principle is that the larger the value, the smaller the output result.

[0062] Based on the above embodiments, as an optional embodiment, a second score is determined according to the maximum value, the minimum value in the target value range, and the performance parameter value, including: Obtain the boundary values ​​of the pre-set evaluation model output results; the boundary values ​​include the maximum boundary value and the minimum boundary value, the maximum boundary value is less than the first score; the minimum boundary value is greater than the third score; By defining the boundary values ​​and the target value range, the key parameter values ​​of the evaluation model are determined. The second score is determined based on the key parameter values ​​and performance parameter values.

[0063] In a specific embodiment, the second score is calculated. The key lies in calculating the critical parameter values ​​in the evaluation model. These critical parameter values ​​are determined based on the boundary values, performance parameter values, and the maximum and minimum values ​​within the target value range of the evaluation model's output.

[0064] Therefore, in one optional embodiment, boundary values ​​for the evaluation model output are preset, including a maximum boundary value and a minimum boundary value, wherein the maximum boundary value is less than the first score. For example, when the first score When the value is 1, the maximum boundary value can be set to 0.999. In a specific embodiment, the maximum boundary value is infinitely close to the first score. .

[0065] Minimum boundary value is greater than the third score For example, when the third score When the value is 0, the minimum boundary value can be set to 0.001. In a specific embodiment, the minimum boundary value is infinitely close to the third score. In fact, the values ​​of the maximum and minimum boundary values ​​can be set according to actual business needs. For scenarios with high precision requirements, the larger the maximum boundary value, the smaller the minimum boundary value.

[0066] Furthermore, key parameter values ​​are calculated based on the maximum and minimum boundary values, and the maximum and minimum values ​​within the target value range. These key parameter values, along with the performance parameter values, are then used to determine the second score.

[0067] Based on the above embodiments, as an optional embodiment, the key parameter values ​​include a first parameter value and a second parameter value; The value of the first parameter is determined based on the boundary value, the maximum value, and the minimum value. In the first model, the value of the first parameter is less than 0, and in the second model, the value of the first parameter is greater than 0. In the first model, the value of the second parameter is determined based on the minimum value, the value of the first parameter, and the minimum boundary value; In the second model, the value of the second parameter is determined based on the maximum value, the value of the first parameter, and the minimum boundary value.

[0068] Specifically, in one optional embodiment, in the first model, the first parameter can be calculated according to formula (3). In the second model, the first parameter can be calculated according to formula (4): (3) (4) in, The first parameter value, For the maximum boundary value, This is the minimum boundary value.

[0069] Referring to formulas (3) and (4), it can be seen that in the specific embodiment, the maximum value is... Greater than the minimum value Therefore, in the first model, the first parameter value Less than 0. In the second model, the first parameter value Greater than 0.

[0070] In another alternative embodiment, in the first model, the second parameter can be calculated according to formula (5). In the second model, the second parameter can be calculated according to formula (6): (5) (6) in, This is the value of the second parameter.

[0071] As can be seen from formula (5), in the first model, the value of the second parameter is... Based on the minimum value First parameter value and minimum boundary value Confirmed. Referring to formula (6), it can be seen that in the first model, the value of the second parameter is... Based on the maximum value First parameter value and minimum boundary value Sure.

[0072] Based on the above embodiments, as an optional embodiment, the second score... The calculation formula is shown in formula (7): (7) Therefore, in another alternative embodiment, if the first score The value can be 1, and the third score is given. The value can be 0. Formula (1) can be expressed as formula (8), and formula (2) can be expressed as formula (9): (8) (9) in, , .

[0073] Therefore, the evaluation model provided in this application embodiment can directly calculate the dimensional scores under different performance index dimensions, thereby determining the performance of the satellite system in different dimensions. That is, through multi-dimensional evaluation, the reliability, real-time performance, and energy efficiency of the intelligent model in the real space environment are verified. For example, it supports rapid task replanning in disaster monitoring, high-precision calculation in multi-satellite collaborative positioning, and real-time identification of unknown space targets, thereby promoting the leapfrog development of satellites from "passive execution" to "autonomous decision-making".

[0074] In one alternative embodiment, the autonomous metrics include energy efficiency, temperature stability, autonomous repair rate, system fault recovery time, peak model memory usage, and fault diagnosis power consumption. Control indicators include mission planning duration, revisit cycle reduction rate, energy utilization rate improvement rate, anomaly handling success rate, and satellite positioning error; On-orbit computing metrics include target detection accuracy, model inference latency, incremental learning decay, model energy efficiency ratio, and unknown target recall rate.

[0075] In a specific embodiment, referring to Table 1, six indicators are defined for the constellation autonomy layer to evaluate the satellite system's performance in six dimensions, five indicators are defined for the collaborative control layer to evaluate the satellite system's performance in five dimensions, and five indicators are also defined for the on-orbit computing layer to evaluate the satellite system's performance in five dimensions. The following will describe each of these specific performance indicators in detail.

[0076] Energy utilization efficiency (EUE) refers to the percentage of energy consumed by the payload and critical subsystems relative to the total available energy, i.e., EUE = (Effective energy consumption / Total available energy) × 100%. It is a key indicator for evaluating how effectively a satellite utilizes its energy supply to meet mission requirements.

[0077] In one optional embodiment, in a specific embodiment where the second score is determined based on the maximum value, the minimum value within the target value range, and the performance parameter value, before calculation, it is first determined whether the performance parameter value is a specified multiple of 10, where the specified multiple is within a preset range (e.g., between 10 and 100). If yes, the second score is calculated based on the performance parameter value; otherwise, the performance parameter value is scaled to meet the requirement of the specified multiple.

[0078] Understandably, some performance parameter values ​​are very small, resulting in very small dimensional scores, which is detrimental to observing system performance. Similarly, some performance parameter values ​​are very large, resulting in very large dimensional scores, which is also detrimental to system performance analysis. Therefore, in specific embodiments, scaling factors can be used to scale performance parameter values ​​to ensure that specified performance indicators for different dimensions can be evaluated on roughly the same order of magnitude.

[0079] For energy utilization efficiency (EUE), in one optional embodiment, the maximum value is... The possible value is 300, and the minimum value is... The value can be 100, meaning the target value range is 100 to 300, and the scaling factor can be 1000. Since a higher EUE corresponds to a higher energy efficiency rating and better system performance, the first model can be chosen to calculate the dimensional scores, thus obtaining a first-dimensional score under the autonomous indicator. .

[0080] Temperature stability index (TC) refers to the technical system that maintains satellite components within a suitable temperature range during the mission cycle through active or passive means. Assuming the temperature fluctuation range is [a, b]℃, the reference temperature is (a+b) / 2 and the fluctuation temperature TC = ±(ba) / 2. If the temperature fluctuation range is [-40, 50]℃, the reference temperature is 5℃ and the fluctuation temperature TC = ±45℃.

[0081] In one alternative embodiment, for energy utilization efficiency (EUE), the maximum value is... The value can be 10000, and the minimum value is... The value can be 4000, meaning the target value range is 4000 to 10000, and the scaling factor can be 1000. In the specific implementation of the actual evaluation, when assessing temperature stability, the input data is the absolute value of TC. For example, if TC = ±45, then the absolute value of TC, |TC|, is 45. The smaller the absolute value of TC, the better the temperature control effect and the better the system performance. Therefore, a second model can be selected to calculate the dimensional score, thereby obtaining a first-dimensional score under the autonomous index. .

[0082] The autonomous repair rate (ARR) is the probability or effectiveness of a satellite automatically detecting, isolating, and restoring its functionality without human intervention from the ground, through its built-in intelligent algorithms, redundant design, or adaptive mechanisms. ARR = (number of successfully autonomously repaired faults / total number of faults) × 100%.

[0083] In one alternative embodiment, for the autonomous repair rate (ARR), the maximum value is... The possible value is 800, the minimum value. The target value can be 400, meaning the target value range is 400 to 800, and the scaling factor can be 1000. Since a smaller ARR indicates stronger self-healing capabilities and better system performance, a second model can be chosen to calculate the dimensional scores, thus obtaining a first-dimensional score under the autonomy metric. .

[0084] System Fault Recovery Time (SFRT) refers to the average time from the occurrence of a fault to the complete restoration of critical system functions to a serviceable state. Generally, SFRT is less than or equal to 30 minutes.

[0085] In one alternative embodiment, the maximum value for System Fault Recovery Time (SFRT) is... The possible value is 3000, the minimum value. The value can be 500, meaning the target value range is 500 to 3000, and the scaling factor can be 100. The smaller the SFRT, the stronger the fault recovery capability. A second model can be selected to calculate the dimensional scores, thus obtaining a first-dimensional score under the autonomous indicator. .

[0086] Peak Model Memory Usage (PMEM) refers to the maximum instantaneous memory usage of a model during inference or training to store input data, intermediate feature maps, weight parameters, and temporary variables. It is the memory value (in megabytes) reached by the model's running process within time T.

[0087] In an alternative embodiment, for peak model memory usage (PMEM), the maximum value is... The value can be 10000, and the minimum value is... The value can be 5000, meaning the target value range is 5000 to 10000, and the scaling factor can be 100. Since a smaller PMEM results in less model memory usage and better system performance, a second model can be chosen to calculate the dimensional scores, thus obtaining a first-dimensional score under the autonomous metric. .

[0088] Diagnostic power consumption (DPC) refers to the additional energy consumed by the system when running fault detection, location and self-healing procedures. Under normal circumstances, DPC does not exceed 10 watts.

[0089] In one alternative embodiment, for fault diagnosis power consumption (DPC), the maximum value is... The value can be 1000, and the minimum value is... The value can be 500, meaning the target value range is 500 to 1000, and the scaling factor can be 100. Since a smaller DPC results in lower diagnostic power consumption and better system performance, a second model can be chosen to calculate the dimensional score, thus obtaining a first-dimensional score under the autonomous metric. .

[0090] Task planning time (TPT) refers to the total time (in minutes) from the start of task triggering (or input) to the generation of a complete and executable planning scheme (or decision sequence) by the system.

[0091] In one alternative embodiment, for Task Planning Time (TPT), the maximum value is... The possible value is 6000, the minimum value. The value can be 3000, meaning the target value range is 3000 to 6000, and the scaling factor can be 100. Since a smaller TPT results in higher task planning efficiency and better system performance, a second model can be chosen to calculate the dimensional score, thus obtaining a second-dimensional score under the control index. .

[0092] The Revisit Period Reduction Rate (RPR) refers to the minimum time interval between two consecutive observations of the same location by a satellite (or satellite constellation) through technical or strategic means, thereby improving the data acquisition frequency and observation timeliness. RPR = ((Original Revisit Period - New Revisit Period) / Original Revisit Period) × 100%.

[0093] In one alternative embodiment, for the revisit period reduction rate (RPR), the maximum value is... The possible value is 300, and the minimum value is... The value can be 100, meaning the target value range is 100 to 300, and the scaling factor can be 1000. Since a higher RPR results in a greater reduction in the revisit period and better system performance, the first model can be chosen to calculate the dimensional score, thereby obtaining a second-dimensional score under the control indicator. .

[0094] Energy efficiency improvement (ES) refers to optimizing the energy consumption of a satellite during its on-orbit operation through technical or management means, in order to extend mission life, increase payload working time, or reduce dependence on solar arrays / batteries, while ensuring that mission performance is not significantly affected. That is, ES = ((new energy utilization rate - original energy utilization rate) / original energy utilization rate) × 100%.

[0095] In one alternative embodiment, for the energy efficiency improvement rate (ES), the maximum value is... The value can be 200, and the minimum value is... The value can be 50, meaning the target value range is 50 to 200, and the scaling factor can be 1000. Since a larger ES value indicates higher energy efficiency and better system performance, the first model can be chosen to calculate the dimensional score, thus obtaining a second-dimensional score under the control index. .

[0096] The success rate of anomaly handling (SREH) refers to the percentage of times that the satellite system successfully identifies, diagnoses and recovers from anomalies within a certain statistical period, out of the total number of anomalies. It reflects the system's autonomous fault tolerance capability and the effectiveness of ground intervention. That is, SREH = (number of anomalies successfully recovered / total number of anomalies) × 100%.

[0097] In one alternative embodiment, for exception handling success rate (SREH), the maximum value is... The possible value is 950, the minimum value. The value can be 800, meaning the target value range is 800 to 950, and the scaling factor can be 1000. Since a larger SREH value results in stronger anomaly handling and better system performance, the first model can be chosen for dimensional score calculation to obtain a second-dimensional score under the control indicator. .

[0098] Positioning error (PE) refers to the deviation (in meters) between the actual position and the theoretical target position of a satellite in a satellite constellation system when maintaining its predetermined orbital configuration or performing cooperative tasks.

[0099] In one alternative embodiment, for satellite positioning error (PE), the maximum value is... The possible value is 950, the minimum value. The value can be 800, meaning the target value range is 800 to 950, and the scaling factor can be 10. Since a smaller PE results in higher positioning accuracy and better system performance, a second model can be chosen to calculate the dimensional score, thus obtaining a second-dimensional score under the control index. .

[0100] For object detection precision (DP), mAP@0.5 is a core metric used to measure model accuracy in object detection tasks. It represents the mean average precision (AP) of all classes when the Intersection over Union (IoU) threshold is 0.5. ,in, Indicates the number of categories. Indicates the first Average accuracy of class detection.

[0101] In one alternative embodiment, for target detection accuracy (DP), the maximum value is... The possible value is 850, the minimum value. The value can be 500, meaning the target value range is 500 to 850, and the scaling factor can be 1000. Since a larger DP (Dynamic Point Count) results in higher detection accuracy from the intelligent model and better system performance, the first model can be chosen for dimensional score calculation to obtain a third-dimensional score under the control index. .

[0102] Model inference latency (MRD) refers to the total time (in milliseconds) from when input data (such as images or text) enters the model until the model outputs a prediction result.

[0103] In one alternative embodiment, for the Model Inference Delay Rate (MRD), the maximum value is... The possible value is 5000, the minimum value. The value can be 200, meaning the target value range is 200 to 5000, and the scaling factor can be 1. Since a smaller MRD results in higher model inference efficiency and better system performance, a second model can be chosen to calculate the dimensional scores, thereby obtaining a third-dimensional score under the control index. .

[0104] Incremental learning decay (ILD) refers to a parameter update constraint mechanism introduced during continuous learning to address the problem of catastrophic forgetting. It balances the learning intensity of new and old knowledge by reducing the update magnitude of parameters related to the old task. ILD generally does not exceed 5%.

[0105] In one alternative embodiment, for incremental learning decay (ILD), the maximum value is... The value can be 200, and the minimum value is... The value can be 50, meaning the target value range is 50 to 200, and the scaling factor can be 1000. Since a smaller ILD results in a lower risk of catastrophic forgetting and better system performance, a second model can be chosen to calculate the dimensional score, thus obtaining a third-dimensional score under the control index. .

[0106] Model Energy Efficiency Ratio (MEER) is a key indicator for measuring the energy efficiency and real-time performance of edge computing devices. It represents the ability to maintain a stable inference speed with limited energy consumption. MEER = throughput (frames per second) / power consumption (W).

[0107] In one alternative embodiment, for the Model Energy Efficiency Ratio (MEER), the maximum value is... The possible value is 300, and the minimum value is... The value can be set to 150, meaning the target value range is 150 to 300, and the scaling factor can be set to 100. Since a larger MEER results in higher model energy efficiency and better system performance, the first model can be chosen to calculate the dimensional score, thus obtaining a third-dimensional score under the control index. .

[0108] Unknown Target Recall (UTRR) measures a model's ability to detect targets that were not seen during training (unknown categories) in an open world or open set scenario. It reflects the model's sensitivity to and generalization to the "unknown". UTRR = number of targets correctly identified as "unknown" / total number of all unknown targets × 100%.

[0109] In one alternative embodiment, for Unknown Target Recall (UTRR), the maximum value is... The possible value is 800, the minimum value. The target value can be 400, meaning the target value range is 400 to 800, and the scaling factor can be 1000. Since a higher UTRR results in higher model energy efficiency and better system performance, the first model can be chosen to calculate the dimensional scores, thus obtaining a third-dimensional score under the control index. .

[0110] Based on the above embodiments, as an optional embodiment, the specified performance indicators include autonomous indicators, control indicators, and on-orbit computation indicators; based on the dimensional scores, it is determined whether the performance of the satellite system to be evaluated meets expectations, including: Assign a first weight to the first dimension score corresponding to the autonomous indicator; assign a second weight to the second dimension score corresponding to the regulation indicator; and assign a third weight to the third dimension score corresponding to the on-orbit calculation indicator; the sum of each first weight, the sum of each second weight, and the sum of each third weight are all 1. Based on the first weight, the scores of the first dimension are weighted and summed to obtain the autonomous layer score; Based on the second weight, the scores of the second dimension are weighted and summed to obtain the control layer score; Based on the third weight, the scores of the third dimension are weighted and summed to obtain the score of the on-orbit computing layer. Based on the scores of the autonomous layer, the control layer, and the orbital computing layer, it is determined whether the performance of the satellite system to be evaluated meets expectations.

[0111] For the constellation autonomous layer, the score is based on the first dimension. First dimension score First dimension score First dimension score First dimension score And the first dimension score Configure the first weight respectively , , , , and ,in, ,and .

[0112] In one optional embodiment, the weights corresponding to the scores of each first dimension can be set sequentially as follows: , , , , , In specific embodiments, the first weight can be adjusted according to actual business needs. For example, when users are more concerned about energy efficiency, the first weight can be appropriately increased. .thus, Figure 2 The scores of the autonomous levels shown are determined according to formula (10): (10) in, Score the autonomous level.

[0113] It should be noted that, in an optional embodiment, to ensure that the data of each dimension are compared at the same data level, the scores of each first dimension in formula (10) are... These are all performance parameter values ​​scaled by a scaling factor.

[0114] For the collaborative regulation layer, the score is the second dimension. Second dimension score Second dimension score Second dimension score Second dimension score Configure the second weight respectively , , , and ,in, ,and .

[0115] In an alternative embodiment, the second weight for each second dimension score can be set to... , , , , Similarly, in specific embodiments, the magnitude of the second weight can be set according to actual business needs, and this application does not impose any limitations on this. Therefore, Figure 2 The control layer score shown can be determined according to formula (11): (11) in, This is the score for the control layer. Similarly, in a specific embodiment, the scores of each second dimension in formula (11) are... These are all performance parameter values ​​scaled by a scaling factor.

[0116] For the on-orbit computing layer, the score is based on the third dimension. Third dimension score Third dimension score Third dimension score and third dimension score Configure the third weight respectively , , , and ,in, and .

[0117] In an optional embodiment, the third weight for each third dimension score can be set to... , , , , Similarly, in specific embodiments, the magnitude of the third weight can be set according to actual business needs, and this application does not impose any limitations on this. Therefore, Figure 2 The on-orbit computation layer score shown can be determined according to formula (12): (12) in, This is the score for the on-orbit computation layer. Similarly, in a specific embodiment, the scores of each third dimension in formula (12) are... These are all performance parameter values ​​scaled by a scaling factor.

[0118] Based on the above calculation formulas (10), (11) and (12), the following can be calculated: Figure 2 The scores for each layer of the satellite constellation corresponding to the three paths shown are as follows, specifically including the scores for the autonomous layer. , Regulatory layer score and on-orbit computation layer score Therefore, based on the three-tiered scoring, it is possible to analyze and determine whether the performance of the satellite system under evaluation meets expectations.

[0119] Based on the above embodiments, as an optional embodiment, determining whether the performance of the satellite system to be evaluated meets expectations is done according to the autonomous layer score, the control layer score, and the orbit calculation layer score, including: Assign a first coefficient to the score of the autonomous layer, a second coefficient to the score of the control layer, and a third coefficient to the score of the orbital calculation layer; the sum of the first, second, and third coefficients is 1. Based on the first, second, and third coefficients, the scores of the autonomous layer, the regulation layer, and the orbit calculation layer are weighted and summed to obtain the comprehensive score. When the overall score is greater than the threshold, the performance of the satellite system being evaluated is determined to meet expectations.

[0120] In a specific embodiment, in order to achieve a comprehensive performance evaluation of the system under multiple influencing factors, scoring analysis is performed on three types of indicators: autonomous indicators, control indicators, and on-orbit computing indicators.

[0121] Specifically, scores are given to the autonomous level. , Regulatory layer score and on-orbit computation layer score Configure the corresponding coefficients respectively. Among them, the autonomous layer score... Configure the first coefficient , regulatory layer score Configure the second coefficient On-orbit computing layer score Configure the third coefficient Therefore, according to formula (13), the scores of each layer are weighted and summed to obtain the comprehensive score of the system to be evaluated: (13) in, The overall score is calculated based on the total score. and In an optional embodiment, the first coefficient It can be set to 0.3, the second coefficient. It can be set to 0.3, the third coefficient. It can be set to 0.4. Similarly, in specific embodiments, each system can set it according to actual business needs, and this application does not limit this.

[0122] In one optional embodiment, obtaining the performance parameter values ​​of the satellite system to be evaluated on a specified performance index includes: Control the satellite system under evaluation to operate under specified environmental parameter conditions; the specified environmental parameter conditions are that the specified performance parameters are within a pre-set specified value range. During operation, performance parameter values ​​are collected in real time.

[0123] In the above embodiments, by defining autonomous indicators such as energy utilization rate, temperature stability index, and autonomous repair rate, control indicators such as mission planning duration, revisit cycle shortening rate, and energy utilization rate improvement rate, as well as on-orbit calculation indicators such as target detection accuracy, model inference delay rate, and incremental learning decay amount, a comprehensive performance evaluation of the satellite system to be evaluated can be achieved in various dimensions.

[0124] However, to ensure that the evaluation method provided in this application can verify whether the satellite system can operate normally under extremely short environmental conditions, verify whether the constellation has rapid response and coordinated scheduling capabilities under any mission, and verify the data processing accuracy and efficiency of the satellite system, in an optional embodiment, the satellite system to be evaluated can be controlled to operate under specified environmental parameter conditions to collect performance parameter values ​​generated under the specified environmental parameters.

[0125] Specifically, at the constellation autonomy layer, a multi-dimensional quantitative evaluation system for the intelligent management capabilities of the satellite system is constructed based on the six indicators provided in the above embodiments. Supported by resource optimization, environmental adaptability, system resilience, and computational efficiency, the objectivity and comparability of the evaluation are ensured through rigorous standard operating conditions and testing methods. First, baseline performance (e.g., raw energy consumption ratio) is established, and extreme operating conditions (-50℃ to +80℃) are simulated in a controlled environment (thermal vacuum / high and low temperature chamber). Second, preset faults (e.g., sensor failure, power short circuit, etc.) are injected, while key parameters are monitored in real time.

[0126] In a specific embodiment of testing energy utilization, thermal control stability, and autonomous fault repair capabilities in a simulated space environment to ensure stable operation of a single satellite under extreme conditions, as an optional embodiment, the specified parameters for controlling the operation of the satellite system may include: ① Energy utilization ≥ 15%; ② Temperature stability index ≤ ±4℃; ③ Autonomous repair rate ≥ 80%; ④ System fault recovery time ≤ 30 minutes; ⑤ Peak model memory usage ≤ 50MB; ⑥ Fault diagnosis power consumption ≤ 10W.

[0127] At the collaborative control layer, the focus is primarily on the effectiveness of multi-satellite collaborative intelligent decision-making. A quantitative evaluation system covering task response, resource optimization, system robustness, and service accuracy is constructed using five indicators. Based on timeliness (emergency planning), coverage capability (revisit cycle), economy (energy-saving scheduling), fault tolerance (anomaly handling), and service accuracy (collaborative positioning), a combination of high-fidelity simulation and physical verification is employed. The model first establishes baseline performance (such as the original revisit cycle and scheduling energy consumption), then injects sudden tasks (disaster monitoring coordinates) and complex anomalies (attitude loss + solar array failure) into the digital twin environment, monitoring the timeliness and resource consumption of the decision chain in real time.

[0128] In a specific embodiment of simulating emergency mission planning, dynamic revisit optimization, and anomaly injection to verify the constellation's rapid response and collaborative scheduling capabilities, as an optional embodiment, the specified parameters for controlling the operation of the satellite system may include: ① mission planning duration < 3 minutes; ② revisit cycle reduction rate ≥ 40%; ③ energy utilization improvement rate ≥ 25%; ④ anomaly handling success rate ≥ 95%; ⑤ satellite positioning error < 50 meters.

[0129] At the on-orbit computing layer, the focus is primarily on the perception accuracy, real-time performance, and adaptability of the system in the space environment. A quantitative evaluation system covering task performance, resource constraints, and open scenarios is constructed using five indicators. Based on detection accuracy (mAP), response speed (latency), continuous evolution (incremental learning), energy efficiency ratio (power consumption / FPS), and scenario generalization (unknown target recall), the model is validated using highly complex datasets and rigorous hardware platforms. The model first establishes baseline performance (e.g., initial mAP for 30 target classes), then injects multi-scale occlusion images and incremental data streams onto an embedded platform, simultaneously monitoring accuracy decay, latency, and power consumption fluctuations.

[0130] In a specific embodiment of target detection, incremental learning, and low-power inference testing based on a satellite-borne model to ensure the accuracy and efficiency of real-time data processing, as an optional embodiment, the specified parameters for controlling the operation of the satellite system may include: ① target detection accuracy mAP@0.5≥0.85; ② model inference latency ≤100ms; ③ incremental learning attenuation <5%; ④ 10FPS@5W; ⑤ unknown target recall rate ≥80%.

[0131] Therefore, the satellite system performance evaluation method provided in this application significantly reduces human intervention through structured indicators and automated testing processes, enabling rapid and standardized evaluation of intelligent models. Simultaneously, by exposing the model's weaknesses under extreme conditions, it promotes algorithm optimization and hardware adaptation, enhancing the satellite system's on-orbit survivability and mission continuity. In practical applications, it is suitable for satellite systems with different orbits, payloads, and mission types, supporting flexible expansion and customized indicator design, and facilitating technology iteration and comparative verification. Furthermore, it provides key verification methods for the engineering implementation of constellation autonomous management, collaborative decision-making, and onboard AI technologies, accelerating the transformation of technology from the laboratory to space applications.

[0132] In the above embodiments, the method for evaluating satellite system performance has been described in detail. This application also provides an embodiment of a satellite system performance evaluation device.

[0133] Figure 4 This is a schematic diagram of the structure of a satellite system performance evaluation device provided in an embodiment of this application, as shown below. Figure 4 As shown, the device includes: The parameter value acquisition module 40 is used to acquire the performance parameter values ​​of the satellite system to be evaluated on a specified performance index; the specified performance index includes at least one of autonomous index, control index and on-orbit calculation index. The value range determination module 41 is used to obtain the target value range of each specified performance index according to the pre-built mapping relationship; the mapping relationship is the correspondence between the performance index and the value range; the value range is the actual range of values ​​that the corresponding performance index can take when the satellite system is operating normally. The dimension score determination module 42 is used to determine the dimension score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index based on the performance parameter value and the target value range. The performance evaluation module 43 is used to determine whether the performance of the satellite system to be evaluated meets expectations based on the dimensional scores.

[0134] Furthermore, the satellite system performance evaluation device provided in this application embodiment also includes: The model acquisition module is used to acquire pre-built evaluation models; the evaluation models include a first model and a second model; the first model is a piecewise function that monotonically increases with respect to the performance index; the second model is a piecewise function that monotonically decreases with respect to the performance index. The correlation determination module is used to determine whether a specified performance indicator is positively correlated with the satellite system performance. If so, the dimension score is determined by the first model based on the performance parameter value and the target value range. If not, the dimension score is determined by the second model based on the performance parameter value and the target value range.

[0135] The first score determination module is used in the first model to determine the first score as the dimension score if the performance parameter value is greater than the maximum value in the target range. The second score determination module is used in the first model to determine the second score based on the maximum value, the minimum value in the target range, and the performance parameter value if the performance parameter value is within the target range; and to use the second score as the dimension score. The third score determination module is used in the first model to determine the third score as the dimension score if the performance parameter value is less than the minimum value; wherein the first score is greater than the second score; and the second score is greater than the third score.

[0136] The third score determination module is also used in the second model to use the third score as the dimension score if the performance parameter value is greater than the maximum value. The second score determination module is also used in the second model to use the second score as the dimension score if the performance parameter value is within the target value range. The first score determination module is also used in the second model to use the first score as the dimension score if the performance parameter value is less than the minimum value.

[0137] The boundary value acquisition module is used to acquire the boundary values ​​of the pre-set evaluation model output results; the boundary values ​​include the maximum boundary value and the minimum boundary value, the maximum boundary value is less than the first score, and the minimum boundary value is greater than the third score; The key parameter value determination module is used to determine the key parameter values ​​of the evaluation model by using boundary values ​​and target value ranges. The second score determination module is also used to determine the second score based on key parameter values ​​and performance parameter values; wherein, the key parameter values ​​include first parameter values ​​and second parameter values; the first parameter value is determined based on boundary values, maximum values, and minimum values, and in the first model, the first parameter value is less than 0, while in the second model, the first parameter value is greater than 0; in the first model, the second parameter value is determined based on the minimum value, the first parameter value, and the minimum boundary value; in the second model, the second parameter value is determined based on the maximum value, the first parameter value, and the minimum boundary value.

[0138] The weight allocation module is used to assign first weights to the first dimension scores corresponding to autonomous indicators; assign second weights to the second dimension scores corresponding to control indicators; and assign third weights to the third dimension scores corresponding to on-orbit calculation indicators; the sum of each first weight, each second weight, and each third weight is 1. The first weighted summation module is used to perform weighted summation on the scores of the first dimension based on the first weight to obtain the autonomous layer score; to perform weighted summation on the scores of the second dimension based on the second weight to obtain the regulation layer score; and to perform weighted summation on the scores of the third dimension based on the third weight to obtain the on-orbit computing layer score. The expectation determination module is used to determine whether the performance of the satellite system to be evaluated meets expectations based on the scores of the autonomous layer, the control layer, and the orbit calculation layer.

[0139] The coefficient allocation module is used to configure a first coefficient for the autonomous layer score, a second coefficient for the control layer score, and a third coefficient for the orbit calculation layer score; the sum of the first, second, and third coefficients is 1. The second weighted summation module is used to perform a weighted summation of the scores of the autonomous layer, the control layer, and the orbit calculation layer based on the first coefficient, the second coefficient, and the third coefficient to obtain a comprehensive score. The threshold judgment module is used to determine that the performance of the satellite system being evaluated meets expectations when the comprehensive score is greater than the threshold.

[0140] The operation control module is used to control the operation of the satellite system under evaluation under specified environmental parameter conditions; the specified environmental parameter conditions are that the specified performance parameters are within a pre-set specified value range; The data acquisition module is used to collect performance parameter values ​​in real time during operation.

[0141] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 5 As shown, the electronic device includes: a memory 50 for storing computer programs; The processor 51 is used to execute computer programs to implement the steps of the satellite system performance evaluation method mentioned in the above embodiments.

[0142] The electronic devices provided in this embodiment may include, but are not limited to, laptops or desktop computers.

[0143] The processor 51 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 51 may be implemented using at least one of the following hardware forms: Digital Signal Processor (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 51 may also include a main processor and a coprocessor. The main processor, also known as the Central Processing Unit (CPU), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 51 may integrate a Graphics Processing Unit (GPU), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 51 may also include an Artificial Intelligence (AI) processor, which is used to handle computational operations related to machine learning.

[0144] The memory 50 may include one or more computer-readable storage media, which may be non-transitory. The memory 50 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In this embodiment, the memory 50 is used to store at least the following computer program 501, which, after being loaded and executed by the processor 51, is capable of implementing the relevant steps of the satellite system performance evaluation method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 50 may also include an operating system 502 and data 503, and the storage method may be temporary or permanent storage. The operating system 502 may include Windows, Unix, Linux, etc. The data 503 may include, but is not limited to, relevant data involved in the satellite system performance evaluation method.

[0145] In some embodiments, the electronic device may further include a display screen 52, an input / output interface 53, a communication interface 54, a power supply 55, and a communication bus 56.

[0146] Those skilled in the art will understand that Figure 5 The structures shown do not constitute a limitation on electronic devices and may include more or fewer components than those shown.

[0147] The electronic device provided in this application includes a memory and a processor. When the processor executes the program stored in the memory, it can implement the satellite system performance evaluation method in the above embodiments.

[0148] It should be noted that although the operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or sequentially, or requiring all illustrated operations to be performed to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system modules and components in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

Claims

1. A method for evaluating the performance of a satellite system, characterized in that, The method includes: Obtain the performance parameter values ​​of the satellite system to be evaluated on a specified performance index; the specified performance index includes at least one of autonomous index, control index, and on-orbit computation index. Based on the pre-constructed mapping relationship, the target value range of each specified performance indicator is obtained; the mapping relationship is the correspondence between the performance indicator and the value range; the value range is the actual range of values ​​that the corresponding performance indicator can take when the satellite system is operating normally. Based on the performance parameter values ​​and the target value range, determine the dimension score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index; Based on the scores of the aforementioned dimensions, it is determined whether the performance of the satellite system to be evaluated meets expectations.

2. The method for evaluating the performance of a satellite system as described in claim 1, characterized in that, The step of determining the dimensional score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index based on the performance parameter value and the target value range includes: Obtain a pre-built evaluation model; the evaluation model includes a first model and a second model; the first model is a piecewise function that monotonically increases with respect to the performance index; the second model is a piecewise function that monotonically decreases with respect to the performance index. Determine whether the specified performance indicators are positively correlated with the satellite system performance; If so, the dimension score is determined using the first model based on the performance parameter value and the target value range; If not, the dimension score is determined using the second model based on the performance parameter values ​​and the target value range.

3. The method for evaluating the performance of a satellite system as described in claim 2, characterized in that, The step of determining the dimension score based on the performance parameter value and the target value range using the first model includes: If the performance parameter value is greater than the maximum value in the target value range, the first score will be used as the dimension score. If the performance parameter value is within the target value range, a second score is determined based on the maximum value, the minimum value within the target value range, and the performance parameter value; and the second score is used as the dimension score. If the performance parameter value is less than the minimum value, the third score is used as the dimension score; wherein the first score is greater than the second score; and the second score is greater than the third score.

4. The method for evaluating the performance of a satellite system as described in claim 3, characterized in that, The step of determining the dimension score based on the performance parameter value and the target value range using the second model includes: If the performance parameter value is greater than the maximum value, the third score will be used as the dimension score. If the performance parameter value is within the target range, the second score will be used as the dimension score. If the performance parameter value is less than the minimum value, the first score is used as the dimension score.

5. The method for evaluating the performance of a satellite system as described in claim 3, characterized in that, The determination of the second score based on the maximum value, the minimum value within the target value range, and the performance parameter value includes: Obtain the boundary values ​​of the pre-set evaluation model output results; the boundary values ​​include a maximum boundary value and a minimum boundary value, wherein the maximum boundary value is less than the first score; and the minimum boundary value is greater than the third score; The key parameter values ​​of the evaluation model are determined by the boundary values ​​and the target value range. The second score is determined based on the key parameter values ​​and the performance parameter values.

6. The method for evaluating the performance of a satellite system as described in claim 5, characterized in that, The key parameter values ​​include the first parameter value and the second parameter value; The first parameter value is determined based on the boundary value, the maximum value, and the minimum value. In the first model, the first parameter value is less than 0, and in the second model, the first parameter value is greater than 0. In the first model, the second parameter value is determined based on the minimum value, the first parameter value, and the minimum boundary value; In the second model, the second parameter value is determined based on the maximum value, the first parameter value, and the minimum boundary value.

7. The method for evaluating the performance of a satellite system as described in claim 1, characterized in that, The autonomous indicators include energy utilization rate, temperature stability index, autonomous repair rate, system fault recovery time, peak model memory usage, and fault diagnosis power consumption. The control indicators include mission planning duration, revisit cycle reduction rate, energy utilization rate improvement rate, anomaly handling success rate, and satellite positioning error. The on-orbit computing metrics include target detection accuracy, model inference latency, incremental learning decay, model energy efficiency ratio, and unknown target recall rate.

8. The method for evaluating the performance of a satellite system as described in claim 1, characterized in that, The specified performance indicators include the autonomous indicators, the control indicators, and the on-orbit computation indicators; determining whether the performance of the satellite system to be evaluated meets expectations based on the dimensional scores includes: First weights are assigned to the first dimension scores corresponding to the autonomous indicators; second weights are assigned to the second dimension scores corresponding to the control indicators; and third weights are assigned to the third dimension scores corresponding to the on-orbit computing indicators; the sum of each first weight, the sum of each second weight, and the sum of each third weight are all 1. Based on the first weight, the scores of the first dimension are weighted and summed to obtain the autonomous layer score; Based on the second weight, the scores of the second dimension are weighted and summed to obtain the control layer score; Based on the third weight, the scores of the third dimension are weighted and summed to obtain the on-orbit computing layer score; Based on the scores of the autonomous layer, the control layer, and the orbit calculation layer, it is determined whether the performance of the satellite system to be evaluated meets expectations.

9. The method for evaluating the performance of a satellite system as described in claim 8, characterized in that, The step of determining whether the performance of the satellite system to be evaluated meets expectations based on the autonomous layer score, the control layer score, and the orbit calculation layer score includes: A first coefficient is configured for the score of the autonomous layer, a second coefficient is configured for the score of the control layer, and a third coefficient is configured for the score of the orbit calculation layer; the sum of the first coefficient, the second coefficient, and the third coefficient is 1. Based on the first coefficient, the second coefficient, and the third coefficient, the scores of the autonomous layer, the control layer, and the orbit calculation layer are weighted and summed to obtain a comprehensive score. When the overall score is greater than the threshold, it is determined that the performance of the satellite system to be evaluated meets expectations.

10. The method for evaluating the performance of a satellite system as described in claim 1, characterized in that, Obtain the performance parameter values ​​of the satellite system to be evaluated on the specified performance indicators, including: The satellite system to be evaluated is controlled to operate under specified environmental parameter conditions; the specified environmental parameter conditions are that the specified performance parameters are within a pre-set specified value range. The performance parameter values ​​are collected in real time during operation.

11. A device for evaluating the performance of a satellite system, characterized in that, The device includes: The parameter value acquisition module is used to acquire the performance parameter values ​​of the satellite system to be evaluated on a specified performance index; the specified performance index includes at least one of autonomous index, control index and on-orbit calculation index. The value range determination module is used to obtain the target value range of each specified performance indicator according to a pre-constructed mapping relationship; the mapping relationship is the correspondence between the performance indicator and the value range; the value range is the actual range of values ​​that the corresponding performance indicator can take when the satellite system is operating normally. The dimension score determination module is used to determine the dimension score of the satellite system to be evaluated on the evaluation dimension corresponding to the specified performance index based on the performance parameter value and the target value range. The performance evaluation module is used to determine whether the performance of the satellite system to be evaluated meets expectations based on the scores of the dimensions.

12. An electronic device comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the satellite system performance evaluation method according to any one of claims 1 to 10.

Images