Adaptive data encoding method, electronic device, and storage medium

Abstract

The application discloses a kind of self-adapting data coding method, electronic equipment and storage medium, the method includes determining task type and task feature vector based on data processing task;Obtain to be processed data, space environment monitoring information and resource constraint information;Determine data feature vector based on to be processed data, determine space environment feature vector based on space environment monitoring information;Determine resource constraint feature vector based on resource constraint information;Determine the constraint information corresponding to data processing task based on task feature vector, data feature vector, space environment feature vector and resource constraint feature vector;Determine the coding granularity of to be processed data based on constraint information;To be processed data is encoded based on coding granularity and constraint information to obtain encoded data, and the encoded data is processed based on task type.This application can adaptively encode data according to space environment, improve the flexibility of data coding and the adaptability to extreme space environment.

Description

Technical Field

[0001] This invention relates to the field of storage technology, specifically to an adaptive data encoding method, electronic device, and storage medium. Background Technology

[0002] In the aerospace field, storage devices not only need to have high-density data storage capabilities, but also need to maintain extremely high reliability under complex space environments (such as high and low temperature cycles, space radiation, and particle impacts). To better store data, it is necessary to encode the data.

[0003] Existing technologies include some adaptive coding methods, which mainly adjust the coding algorithm according to the data type. However, these algorithms are all fixed-granularity coding algorithms, where the granularity of coding (i.e., the size of the coded data block) is fixed.

[0004] However, in safe environments with low radiation intensity, low-granularity coding can generate excessive redundant data, wasting valuable storage space and transmission bandwidth; while in environments with strong radiation or extreme temperatures, fixed-granularity coding may not be sufficient to correct sudden large numbers of bit errors, leading to the loss of critical data. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this application provides an adaptive data encoding method, electronic device, and storage medium. By determining the encoding granularity of the data to be processed based on constraint information, it can perform adaptive data encoding according to the spatial environment, thereby improving the flexibility of data encoding and its adaptability to extreme spatial environments.

[0006] To address the above problems, the present invention provides the following technical solution: In a first aspect, embodiments of this application provide an adaptive data encoding method, comprising: determining a task type and a task feature vector based on the current data processing task, wherein the task type is uploading data or storing data; Acquire the data to be processed, space environment monitoring information, and resource constraint information corresponding to the data processing task; Determine the data feature vector based on the data to be processed. Determine the space environment feature vector based on the aforementioned space environment monitoring information; Determine the resource constraint feature vector based on the aforementioned resource constraint information; The constraint information corresponding to the data processing task is determined based on the task feature vector, the data feature vector, the spatial environment feature vector, and the resource constraint feature vector. The encoding granularity of the data to be processed is determined based on the constraint information; The data to be processed is encoded based on the encoding granularity and the constraint information to obtain encoded data, and the encoded data is processed based on the task type.

[0007] Optionally, the task feature vector includes at least one feature parameter selected from task security level, task real-time performance level, and task importance level; and / or The data feature vector includes at least one of the following feature parameters: data type parameter, access frequency, lifespan, and security level; and / or The spatial environment feature vector includes at least one feature parameter selected from equipment temperature, radiation intensity, channel bit error rate, and link signal-to-noise ratio; and / or The resource constraint feature vector includes at least one of the following feature parameters: processor computing power margin, storage capacity margin, power supply margin, and available bandwidth.

[0008] Optionally, determining the encoding granularity of the data to be processed based on the constraint information includes: The initial coding granularity is determined based on the data feature vector in the constraint information and multiple preset coding granularities; The environmental granularity adjustment amount is determined based on the aforementioned spatial environment feature vector; The task granularity adjustment amount is determined based on the task feature vector; The resource constraint adjustment amount is determined based on the aforementioned resource constraint feature vector; The final coding granularity is determined based on the initial coding granularity, the environment granularity adjustment amount, the task granularity adjustment amount, and the resource constraint adjustment amount.

[0009] Optionally, the space environment feature vector includes device temperature, radiation intensity, and single-event turnover rate, and the step of determining the environmental granularity adjustment amount based on the space environment feature vector includes: The environmental particle size adjustment amount is determined based on the ratio of the device temperature to a preset temperature threshold and the ratio of the radiation intensity to a preset radiation intensity threshold.

[0010] Optionally, the task feature vector includes a task real-time level, and determining the task granularity adjustment amount based on the task feature vector includes: The task granularity adjustment amount is determined based on the task real-time level in the task feature vector, wherein the task granularity adjustment amount is negatively correlated with the task real-time level.

[0011] Optionally, the resource constraint feature vector includes processor computing power margin, and determining the resource constraint adjustment amount based on the resource constraint feature vector includes: The resource constraint adjustment amount is determined based on the processor computing power margin in the resource constraint feature vector, wherein the resource constraint adjustment amount is positively correlated with the processor computing power margin.

[0012] Optionally, the step of encoding the data to be processed based on the encoding granularity and the constraint information to obtain encoded data, and processing the encoded data based on the task type, includes: Based on the encoding granularity, the data to be processed is divided into multiple data blocks to be processed; The target encoding method for the data block to be processed is determined based on the encoding granularity and the constraint information. Each of the data blocks to be processed is encoded using the target encoding method to obtain multiple encoded data blocks; The encoded data is obtained by concatenating multiple encoded data blocks, and the encoded data is processed based on the task type.

[0013] Optionally, determining the target encoding method for the data block to be processed based on the encoding granularity and the constraint information includes: Determine the constraint feature vector based on the constraint information; A performance quantization feature vector is determined for each of the preset multiple encoding methods corresponding to the encoding granularity. The performance quantization feature vector includes compression ratio, encryption strength, error correction capability, computing power consumption, and encoding latency. Based on the constraint feature vector and the performance quantization feature vector, one of the multiple encoding methods is selected as the target encoding method.

[0014] Secondly, embodiments of this application provide an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the adaptive data encoding method as described in the first aspect.

[0015] Thirdly, embodiments of this application provide a computer-readable storage medium storing an executable program, which is executed by a processor to implement the adaptive data encoding method as described in the first aspect.

[0016] This application provides an adaptive data encoding method, electronic device, and storage medium. By determining the encoding granularity of the data to be processed based on constraint information, this application can perform adaptive data encoding according to the space environment, thereby improving the flexibility of data encoding and its adaptability to extreme space environments. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the adaptive data encoding method provided in the embodiments of this application.

[0018] Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0019] Figure 3 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "multiple" means two or more, unless otherwise explicitly specified.

[0022] This application provides an adaptive data encoding method, electronic device, and storage medium. By determining the encoding granularity of the data to be processed based on constraint information, it can perform adaptive data encoding according to the spatial environment, thereby improving the flexibility of data encoding and its adaptability to extreme spatial environments.

[0023] This application pertains to the next-generation information network industry and should be classified as H04N19.

[0024] The adaptive data encoding method provided in this application will be described in detail below with reference to the accompanying drawings.

[0025] Please see Figure 1 , Figure 1 This is a flowchart illustrating the adaptive data encoding method provided in an embodiment of this application. Figure 1 As shown, the adaptive data encoding method includes steps S100 to S500.

[0026] Step S100: Determine the task type and task feature vector based on the current data processing task.

[0027] The task type is either uploading data or storing data.

[0028] Optionally, the task feature vector includes at least one feature parameter selected from the task security level, task real-time level, and task importance level.

[0029] Optionally, in response to receiving the execution instruction of the data processing task, the execution instruction is parsed to obtain the task type and all feature parameters in the task feature vector.

[0030] Step S200: Obtain the data to be processed, space environment monitoring information, and resource constraint information corresponding to the data processing task.

[0031] Optionally, the space environment monitoring information includes at least one of the following: device temperature of the onboard storage device, radiation intensity, channel bit error rate, and link signal-to-noise ratio.

[0032] Optionally, the resource constraint information includes at least one of processor computing power margin, storage capacity margin, power supply margin, and available bandwidth.

[0033] Step S300: Determine the data feature vector based on the data to be processed.

[0034] Optionally, the data feature vector includes at least one of the following feature parameters: data type parameter, access frequency, lifespan, and security level.

[0035] The data lifecycle refers to the time it takes for data to go through the entire process from creation, modification, publication and utilization to archiving or destruction.

[0036] Optionally, the data to be processed is parsed to determine its data type, which can be text, image, table, or audio. Then, the data type parameters are determined based on the data type of the data to be processed and the preset mapping relationship between data types and data type parameters.

[0037] Optionally, each feature parameter is normalized, and then the data feature vector is determined based on the normalized feature parameters.

[0038] For example, the difference between a feature parameter and its corresponding preset minimum value is divided by the preset normalized value range corresponding to the feature parameter to obtain the normalized feature parameter.

[0039] Step S400: Determine the space environment feature vector based on space environment monitoring information.

[0040] Optionally, the space environment monitoring information includes monitoring data from environmental monitoring equipment, such as equipment temperature and radiation intensity.

[0041] Optionally, the space environment monitoring information may also include the operating logs of the storage devices.

[0042] Optionally, each feature parameter is normalized, and then the spatial environment feature vector is determined based on the normalized feature parameters, as described above.

[0043] Optionally, the spatial environment feature vector includes at least one feature parameter among equipment temperature, radiation intensity, channel bit error rate, and link signal-to-noise ratio.

[0044] Optionally, based on the mission type and the preset correspondence between mission type and feature parameter weights, each initial feature parameter is multiplied by its corresponding weight to obtain weighted feature parameters. Finally, the space environment feature vector is determined based on all weighted feature parameters. This dynamic weight adaptation method can meet the core requirements of different aerospace missions.

[0045] For example, when the task type is uploading data, the channel bit error rate has the highest weight. When the task type is storing data, the radiation intensity has the highest weight.

[0046] Step S500: Determine the resource constraint feature vector based on resource constraint information.

[0047] Optionally, the resource constraint information includes at least one of processor computing power margin, storage capacity margin, power supply margin, and available bandwidth.

[0048] Optionally, the resource constraint feature vector includes at least one of the following feature parameters: processor computing power margin, storage capacity margin, power supply margin, and available bandwidth.

[0049] Optionally, each feature parameter is normalized, and then the resource constraint feature vector is determined based on the normalized feature parameters, as described above.

[0050] Step S600: Determine the constraint information corresponding to the data processing task based on the task feature vector, data feature vector, spatial environment feature vector, and resource constraint feature vector.

[0051] Among them, the task feature vector, data feature vector, spatial environment feature vector, and resource constraint feature vector have the same dimension.

[0052] Optionally, the constraint information includes constraint feature vectors.

[0053] Optionally, a preset method can be used to fuse task feature vectors, data feature vectors, spatial environment feature vectors, and resource constraint feature vectors into constraint information.

[0054] Optionally, the task feature vector, data feature vector, spatial environment feature vector, and resource constraint feature vector can be concatenated to obtain constraint information.

[0055] Step S700: Determine the encoding granularity of the data to be processed based on the constraint information.

[0056] In this context, coding granularity refers to the size of the smallest independent data unit in a single coding operation, which is also the size of the coded data block. For example, the coding granularity can be 512B, 2KB, 4KB, or 16KB. Variable coding granularity essentially represents a trade-off between reliability and coding efficiency. Smaller coding granularity results in stronger independence of the coded data blocks and a smaller impact range from bit errors. Larger coding granularity allows the coding algorithm to fully utilize data features, leading to a higher theoretical performance ceiling.

[0057] Specifically, for error correction coding algorithms, the smaller the coding granularity, the higher the proportion of correctable errors per block under the same error correction coding algorithm, thus improving the method's resistance to single-event upsets and burst errors. In extreme radiation environments, continuous errors caused by single-event upsets can be confined to a single small block, making it difficult to exceed the error correction limit and thus preventing the entire data block from failing. Therefore, it is necessary to reduce the coding granularity in strong radiation environments. The larger the coding granularity, the closer the error correction performance of the error correction coding algorithm is to the Shannon limit, and the higher the coding efficiency.

[0058] For compression coding algorithms, the larger the coding granularity, the more data correlations and global repetition patterns the algorithm can capture, resulting in higher efficiency for dictionary compression and entropy coding, and a greater compression ratio. For example, for space remote sensing images, a large coding granularity allows the algorithm to utilize the features of the entire data segment, achieving a compression ratio far higher than small-block segmentation. This maximizes the saving of valuable onboard storage capacity and satellite-to-ground link bandwidth, so the coding granularity needs to be increased in stable environments. Conversely, the smaller the coding granularity, the fewer data correlations the compression coding algorithm can capture, and the lower the compression efficiency.

[0059] For encryption algorithms, the smaller the encoding granularity, the smaller the leakage range and the higher the security. Small data blocks can also achieve real-time streaming encryption without waiting for large blocks of data to be ready, making it suitable for high-real-time data processing tasks. However, this results in high computational overhead. Conversely, the larger the encoding granularity, the more the fixed encryption overhead can be distributed across massive amounts of data, reducing the computational overhead per unit of data. However, this leads to a larger leakage range and lower security.

[0060] Optionally, step S700 includes steps S710 to S750.

[0061] Step S710: Determine the initial coding granularity based on the data feature vector in the constraint information and multiple preset coding granularities.

[0062] Optionally, a data feature matching template corresponding to each coding granularity is determined in advance based on historical data, the cosine similarity between the data feature vector and each data feature matching template is calculated, and then the coding granularity with the highest cosine similarity is determined as the initial coding granularity.

[0063] Step S720: Determine the environmental granularity adjustment amount based on the spatial environment feature vector.

[0064] Optionally, the environmental particle size adjustment amount is determined based on the ratio of the equipment temperature to a preset temperature threshold and the ratio of the radiation intensity to a preset radiation intensity threshold.

[0065] Optionally, a space environment severity index is determined based on the ratio of equipment temperature to a preset temperature threshold and the ratio of radiation intensity to a preset radiation intensity threshold. When the space environment severity index is less than 1, the space environment is considered safe, and the environmental granularity adjustment amount is linearly negatively correlated with the space environment severity index. When the space environment severity index is not less than 1, the space environment is considered unsafe, and the environmental granularity adjustment amount is exponentially negatively correlated with the space environment severity index. In this way, the calculation of the environmental granularity adjustment amount distinguishes between the linear adjustment zone and the nonlinear emergency zone, which can better reflect the situation of extreme environments.

[0066] For example, the formula for calculating the space environment severity index is: , in, This indicates the severity index of the space environment. Indicates the first calculated coefficient. This represents the second calculation coefficient. This indicates the current device temperature. Indicates the preset temperature threshold. This indicates the current radiation intensity. This indicates the preset radiation intensity threshold. and All are preset values ​​and are positive numbers.

[0067] For example, the formula for calculating the environmental particle size adjustment amount is: , in, Indicates the amount of environmental particle size adjustment. This represents the preset environmental granularity adjustment amount based on the calculated value. It is a positive number. This represents the preset environmental granularity adjustment coefficient. It is a positive number. It represents the base of the natural logarithm.

[0068] Step S730: Determine the task granularity adjustment amount based on the task feature vector.

[0069] Optionally, the task feature vector includes a task real-time performance level. The task granularity adjustment is determined based on the task real-time performance level in the task feature vector, where the task granularity adjustment is negatively correlated with the task real-time performance level. A smaller task real-time performance level indicates a higher real-time requirement.

[0070] Optionally, the feature parameters in the task feature vector are weighted and summed using a preset combination of weight coefficients corresponding to the task feature vector, and the weighted sum is multiplied by... This yields the task granularity adjustment amount. The weight coefficients in the preset weight coefficient combination corresponding to the task feature vector are negative.

[0071] Step S740: Determine the resource constraint adjustment amount based on the resource constraint feature vector.

[0072] Optionally, the resource constraint adjustment amount is determined based on the processor computing power margin in the resource constraint feature vector, wherein the resource constraint adjustment amount is positively correlated with the processor computing power margin.

[0073] Optionally, the feature parameters in the resource constraint feature vector are weighted and summed using a preset combination of weight coefficients corresponding to the resource constraint feature vector, and the weighted sum is multiplied by... This yields the resource constraint feature vector. The weight coefficients in the preset weight coefficient combination corresponding to the resource constraint feature vector are positive numbers.

[0074] Step S750: Determine the final coding granularity based on the initial coding granularity, environment granularity adjustment amount, task granularity adjustment amount, and resource constraint adjustment amount.

[0075] Optionally, the initial coding granularity is supplemented with the environment granularity adjustment, task granularity adjustment, and resource constraint adjustment to obtain the final coding granularity.

[0076] Step S800: Encode the data to be processed based on the encoding granularity and constraint information to obtain encoded data, and process the encoded data based on the task type.

[0077] Optionally, step S800 includes steps S810 to S840.

[0078] Step S810: Divide the data to be processed into multiple data blocks based on the encoding granularity.

[0079] Step S820: Determine the target encoding method for the data block to be processed based on the encoding granularity and constraint information.

[0080] Optionally, step S820 includes steps S821 to S828.

[0081] Step S821: Determine the constraint feature vector based on the constraint information.

[0082] Step S822: Determine the performance quantization feature vector corresponding to the coding granularity of each of the preset multiple coding methods.

[0083] The performance quantification feature vector includes at least one performance parameter among compression ratio, encryption strength, error correction capability, computing power consumption, and encoding latency.

[0084] Optionally, the performance parameters in the performance quantization feature vector are all normalized.

[0085] Optionally, performance tests are performed on each of the multiple encoding methods at multiple preset encoding granularities to obtain a performance quantization feature vector for each encoding method. Then, for the current encoding granularity, interpolation calculations are performed based on the performance quantization feature vectors of each encoding method at multiple encoding granularities to obtain a performance quantization feature vector for each encoding method corresponding to the current encoding granularity.

[0086] Optionally, the interpolation calculation methods include linear interpolation, polynomial interpolation, and nearest neighbor interpolation.

[0087] Step S823: Select one of the multiple encoding methods as the target encoding method based on the constraint feature vector and the performance quantization feature vector.

[0088] Optionally, constraints are determined based on resource constraint feature vectors in the constraint feature vectors, and a multi-objective optimization algorithm is used to select one of multiple encoding methods as the target encoding method based on the constraints and performance quantization feature vectors.

[0089] For example, the objective function in a multi-objective optimization algorithm is: , in, Describe the objective optimization function. , , , and The dynamic weights of each performance parameter in the performance quantization feature vector are all positive numbers and satisfy the following conditions: . Indicates the first Performance parameters of each encoding method Indicates the first Compression ratio of each encoding method Indicates the first The encryption strength of each encoding method, Indicates the first The error correction capability of each encoding method Indicates the first The computational power consumption of each encoding method Indicates the first The encoding delay of each encoding method.

[0090] Optionally, , , , and This is the default value.

[0091] Optionally, , , , and It is determined by an artificial intelligence model based on historical coding performance data. , , , and It can be updated dynamically.

[0092] The core logic of the objective optimization function is to maximize the compression efficiency, security, and error correction capability of the encoding, while minimizing computational power consumption and encoding latency, achieving a balanced optimal solution for multiple objectives. In this way, the target encoding method can be selected based on the usage scenario and environment of the satellite storage device, thereby improving encoding efficiency.

[0093] Optionally, multi-objective optimization algorithms include particle swarm optimization, genetic algorithms, and ant colony optimization, among others. The goal is to maximize the objective function.

[0094] Step S830: Encode each data block to be processed using the target encoding method to obtain multiple encoded data blocks.

[0095] Step S840: Concatenate multiple encoded data blocks to obtain encoded data, and process the encoded data based on the task type.

[0096] In summary, the adaptive data encoding method provided in this application has the following advantages: 1. By determining the encoding granularity of the data to be processed based on constraint information, adaptive data encoding can be performed according to the spatial environment, which improves the flexibility of data encoding and its adaptability to extreme spatial environments.

[0097] 2. By distinguishing between the linear adjustment zone and the nonlinear emergency zone in the calculation of environmental granularity adjustment, the situation of extreme environments can be better reflected.

[0098] 3. By using the target optimization function, the target encoding method can be selected according to the usage scenario and environment of the storage device on the satellite, thereby improving encoding efficiency.

[0099] All parameters used in the calculations in this application have been normalized. For example, the Min-Max normalization algorithm is used to normalize all parameters used in the calculations.

[0100] Optionally, the encoded data is processed based on the task type, including: classifying the encoded data, and processing the encoded and classified data based on the task type.

[0101] Currently, the sensor data is first initially classified using a pre-trained classification model. Then, a general drift detection algorithm is used to monitor changes in the statistical distribution of the sensor data. When an anomaly in the statistical distribution is detected, it is determined that concept drift has occurred, triggering an update of the classification model. This leads to adjustments in the data classification strategy and the completion of data classification and storage.

[0102] However, general drift detection algorithms only focus on changes in the statistical distribution of sensor data. They are prone to misjudging changes in the statistical distribution of sensor data caused by changes in the normal operating conditions of the satellite (such as load changes) as fault drift. At this time, the health status of the satellite remains unchanged, which will frequently trigger invalid classification model updates. This not only increases the system's computing power consumption, but also leads to confusion in the data classification strategy and reduces the accuracy of data classification.

[0103] Optionally, the encoded data is classified, including steps S910 to S980.

[0104] Step S910: Obtain historical operational data and component association information of the satellite.

[0105] The component association information is used to represent the impact of the operating conditions of each component in the satellite on other components. The component association information is related to the actual physical structure of the satellite.

[0106] Optionally, the historical operating data includes historical sensing data from at least one sensor of the satellite in a fault-free operating state, with each sensor corresponding to the monitoring of the operating status of at least one component.

[0107] Optionally, historical working data is manually selected.

[0108] Optionally, at least one sensor includes a temperature sensor, a vibration sensor, a sound sensor, and a Hall sensor, etc.

[0109] For example, since wear on the motor rotor can cause an increase in bearing vibration amplitude, the motor is associated with the bearing in the component association information.

[0110] Step S920: Determine the satellite's operational condition drift propagation reference directed graph based on historical operating data and component association information.

[0111] In this way, the directed graph of the operational condition drift propagation reference can be linked to the actual physical structure of the satellite, thereby improving the accuracy of data classification.

[0112] Optionally, step S920 includes steps S921 to S923.

[0113] Step S921: Determine the feature tensors of each component of the satellite corresponding to different combinations of operating parameters based on historical working data.

[0114] Optionally, the feature tensor of a component includes feature components from historical sensing data of all sensors used to monitor the component. To unify the dimensionality, the number of feature components in the component's feature tensor is equal to the number of all sensor types; if a type of sensor is not used to monitor the component, the feature component corresponding to that type of sensor is set to 0.

[0115] Optionally, time-domain and / or frequency-domain feature values ​​of historical sensing data of a class of sensors are calculated, and feature components of the class of sensors are determined based on the normalized time-domain and / or frequency-domain feature values.

[0116] Optionally, time-domain features include peak value, mean value, kurtosis, etc.

[0117] Optionally, frequency domain eigenvalues ​​include characteristic frequencies, fundamental amplitude, harmonic amplitude, etc.

[0118] Optionally, the combination of operating parameters includes operating parameters such as satellite load, power, and rotational speed.

[0119] Step S922: Map each component of the satellite to a node in the directed graph of the operating condition drift propagation reference, and determine the directed edges between adjacent nodes based on the component association information.

[0120] For example, if wear on the motor rotor causes an increase in the bearing vibration amplitude, then the directed edge is determined to point from the node representing the motor to the node representing the bearing.

[0121] Step S923: Determine the weight matrix of each directed edge based on historical working data, thereby obtaining the reference directed graph for working condition drift propagation.

[0122] Optionally, step S923 includes steps S9231 to S9232.

[0123] Step S9231: Based on the sensor data of each component of the satellite corresponding to different combinations of operating parameters in the historical working data, determine the operating condition drift propagation time and operating condition drift propagation probability for each directed edge corresponding to different combinations of operating parameters.

[0124] Under a certain combination of operating parameters, when one of the two associated components changes its operating state, the sensing data of the two components will change sequentially, such as peak values ​​or valley values ​​appearing one after the other.

[0125] Optionally, for the sensing data of the components corresponding to the two nodes connected by each directed edge, two sets of sensing data within a first time window are determined. If the sampling time intervals of the two sets of sensing data are different, the sensing signal sequence with the smaller sampling interval is sampled to update the sensing signal sequence, ensuring that the number of numerical points in the two sensing signal sequences is the same. The duration of the first time window is a second preset duration. Then, the number of sampling intervals that maximizes the cross-correlation sequence of the two sensing signal sequences is calculated, and this number of sampling intervals is multiplied by the sampling time interval of the sensing data to obtain the working condition drift propagation time corresponding to the directed edge.

[0126] Optionally, a baseline value of the sensing data of a type of sensor is calculated, and the value points of the sensing data are detected in chronological order. When the value point of the sensing data is greater than the baseline value plus a drift threshold or less than the baseline value plus a drift threshold, the time at which the sensing data drift occurs is determined.

[0127] Optionally, the drift propagation time of multiple operating conditions corresponding to the directed edge is calculated by using a sliding first time window, and the average time of the drift propagation time of multiple operating conditions is determined as the final drift propagation time of the directed edge.

[0128] Optionally, under a combination of operating parameters, the drift time of the sensor data of each sensor is determined according to the preset drift threshold of the sensor data of each type of sensor corresponding to the combination of operating parameters, thereby determining the drift time of the node corresponding to the component monitored by the sensor, and then calculating the operating drift propagation time between two nodes connected by a directed edge based on the drift time of each node.

[0129] Optionally, the drift time of each node within the third time window is counted using the above method, where the duration of the third time window is a third preset duration. Then, the number of working condition drifts corresponding to each node is determined based on the drift time of each node. Specifically, when multiple drift times of a node occur within the fourth time window after the drift time of a node, or when no drift time of a node occurs, a working condition drift is determined to have occurred. Here, the start time of the fourth time window is the earliest drift time of a node, and the duration is the fourth preset duration. When a drift time of a node occurs outside the fourth time window after the drift time of a node, the working condition drift count is incremented by 1. This process is repeated to determine the number of working condition drifts corresponding to each node. The number of working condition drifts may be 0. Then, for each pair of nodes connected by a directed edge, the number of working condition drifts corresponding to the endpoint node of the directed edge is divided by the number of working condition drifts corresponding to the starting node of the directed edge to obtain the working condition drift propagation probability corresponding to the starting node of the directed edge.

[0130] Step S9232: Based on the working condition drift propagation time and working condition drift propagation probability of each directed edge corresponding to different working condition parameter combinations, determine the weight matrix of the directed edge corresponding to different working condition parameter combinations, thereby obtaining the working condition drift propagation reference directed graph.

[0131] Optionally, the weight matrix includes the condition drift propagation time and condition drift propagation probability.

[0132] Step S993: Obtain the combination of satellite sensor data and actual operating parameters within the first time period.

[0133] Optionally, the sensing data includes sensing data from at least one sensor of the satellite.

[0134] The duration of the first time period is the first preset duration, and the end time is the current time.

[0135] Step S994: ​​Determine the actual directed graph of operating condition drift propagation based on the actual combination of operating condition parameters and the reference directed graph of operating condition drift propagation.

[0136] The directed graph of actual operating condition drift propagation is used to reflect the ideal operating condition drift of satellite sensor data operating with actual operating condition parameter combinations.

[0137] Optionally, step S994 includes steps S9941 to S9942.

[0138] Step S9941: Based on the feature tensors of each node in the directed graph of the working condition drift propagation reference, determine the reference feature tensor of the actual working condition parameter combination corresponding to the node.

[0139] Specifically, the feature tensor corresponding to the actual working condition parameter combination is selected from the feature tensors corresponding to different working condition parameter combinations for each node and determined as the benchmark feature tensor.

[0140] Step S9942: Based on the actual working condition parameter combination and the weight matrix of each directed edge corresponding to different working condition parameter combinations, determine the weight matrix of the directed edge corresponding to the actual working condition parameter combination, thereby obtaining the actual directed graph of working condition drift propagation.

[0141] Specifically, the weight matrix corresponding to the actual working condition parameter combination is selected from the weight matrix corresponding to different working condition parameter combinations for each directed edge, thereby obtaining the actual directed graph of working condition drift propagation.

[0142] In this way, for satellites with frequently fluctuating operating parameters (such as CNC machine tools), the directed graph of operating condition drift propagation can dynamically adapt to the actual operating conditions of the satellite, thereby improving the accuracy of data classification.

[0143] Step S995: Determine the actual operating condition directed graph of the satellite based on sensor data, actual operating condition parameter combinations, and component association information.

[0144] Optionally, step S995 includes steps S9951 to S9953.

[0145] Step S9951: Determine the actual feature tensor of each component of the satellite based on the sensor data.

[0146] Step S9952: Map each component of the satellite to a node in the actual operating condition directed graph, and determine the directed edges between all nodes based on the component association information.

[0147] Step S9953: Determine the weight matrix of each directed edge based on the combination of sensor data and actual operating condition parameters, thereby obtaining the directed graph of the actual operating condition.

[0148] The content of step S995 is the same as that of step S920.

[0149] Step S960: Determine the satellite's operational drift degree based on the directed graph of operational drift propagation and the actual operational drift graph.

[0150] The actual operating condition directed graph is used to reflect the superposition of real operating condition drift and possible fault drift in the sensor data of satellites operating with actual operating condition parameter combinations.

[0151] Optionally, step S960 includes steps S961 to S963.

[0152] Step S961: Calculate the working condition deviation of the feature tensors corresponding to two nodes at the same position in the actual working condition directed graph and the actual working condition directed graph.

[0153] Optionally, the cosine similarity of the feature tensors corresponding to two nodes at the same position in the actual working condition drift propagation directed graph and the actual working condition directed graph is calculated, and the cosine similarity is determined as the working condition deviation.

[0154] Optionally, the Euclidean distance between the feature tensors corresponding to two nodes at the same position in the actual working condition drift propagation directed graph and the actual working condition directed graph is calculated, and this Euclidean distance is determined as the working condition deviation.

[0155] Optionally, the magnitude of the subtraction vector of the feature tensor corresponding to two nodes at the same position in the actual working condition drift propagation directed graph and the actual working condition directed graph is calculated, and the magnitude of the subtraction vector is determined as the working condition deviation.

[0156] Step S962: Calculate the propagation deviation of the weight matrices corresponding to the two directed edges at the same position in the actual working condition drift propagation directed graph and the actual working condition directed graph.

[0157] Optionally, a weighted calculation method is used to calculate the propagation deviation corresponding to the two directed edges based on the time difference and probability difference of the operation condition drift propagation in the weight matrix corresponding to the two directed edges at the same position in the actual operation condition directed graph and the actual operation condition directed graph.

[0158] For example, the propagation deviations corresponding to the two directed edges are obtained by multiplying the first preset calculation coefficient by the time difference of the drift propagation under the operating conditions and adding the second preset calculation coefficient by the probability difference of the drift propagation under the operating conditions.

[0159] Step S963: Determine the satellite's operational drift based on operational deviation and propagation deviation.

[0160] Among them, operating condition deviation is used to measure the deviation between the actual working state and the theoretical working state of a component, and propagation deviation is used to measure the degree of state correlation between two related components.

[0161] If the operating condition deviation is large, it indicates that the component may be faulty. If the propagation deviation is large, it indicates that the connection between the two related components may be faulty, or that at least one of the components itself may be faulty.

[0162] By combining the working status of the components themselves with the actual physical structure of the satellite, the operational drift of the satellite can be determined, thereby improving the accuracy of the operational drift.

[0163] Optionally, the operating condition drift can be obtained by multiplying the third preset calculation coefficient by the average of all operating condition deviations and adding the fourth preset calculation coefficient by the average of all propagation deviations.

[0164] Step S970: Determine the drift type of the sensor data based on the operating condition drift.

[0165] Optionally, step S970 includes steps S971 to S975.

[0166] Step S971: Obtain historical operating condition deviation data.

[0167] The historical operating condition deviation data includes the satellite's operating condition drift calculated using the above method up to the current time.

[0168] Step S972: Determine the second and third judgment thresholds based on the preset first judgment threshold and historical operating condition deviation data.

[0169] Optionally, the standard deviation of the working condition deviation of the feature tensors corresponding to two nodes at the same position in multiple sets of historical working condition deviation data is calculated, and the first judgment threshold is added to the first threshold calculation coefficient and multiplied by the standard deviation to obtain the second judgment threshold. The first judgment threshold is added to the second threshold calculation coefficient and multiplied by the standard deviation to obtain the third judgment threshold, wherein the first threshold calculation coefficient is greater than the second threshold calculation coefficient.

[0170] For example, the first threshold calculation coefficient is 1, 1.5, or 2, etc.

[0171] For example, the second threshold calculation coefficient is 0.8, 0.5, or 0.4, etc.

[0172] When a satellite malfunctions, the impact of the malfunction on its operating conditions is generally greater than the impact of normal operating parameters. Therefore, the greater the operating condition drift, the more likely the drift in the sensor data is fault drift.

[0173] Step S973: When the operating condition drift is greater than the second judgment threshold, determine that the drift type of the sensor data is fault drift.

[0174] Step S974: When the operating condition drift is not greater than the second judgment threshold and is greater than the third judgment threshold, the drift type of the sensor data is determined to be operating condition drift.

[0175] Step S975: When the operating condition drift is not greater than the third judgment threshold, determine the drift type of the sensor data as no drift.

[0176] By determining the second and third judgment thresholds as described above, we can adapt to the differences in historical operating conditions of different satellites, avoid misjudgments caused by fixed judgment thresholds, and thus accurately determine the drift type.

[0177] Optionally, the judgment threshold may not be updated. In this case, step S970 includes steps S976 to S978.

[0178] Step S976: When the operating condition drift is greater than the preset first judgment threshold, determine that the drift type of the sensor data is fault drift.

[0179] Step S977: When the operating condition drift is not greater than the first judgment threshold and is greater than the preset fourth judgment threshold, the drift type of the sensor data is determined to be operating condition drift.

[0180] Among them, the first judgment threshold is greater than the fourth judgment threshold.

[0181] Step S978: When the operating condition drift is not greater than the preset fifth judgment threshold, the drift type of the sensor data is determined to be no drift.

[0182] Among them, the fifth judgment threshold is greater than the fourth judgment threshold.

[0183] The two implementation methods described above can be flexibly adapted to different judgment speed requirements, balancing accuracy and efficiency.

[0184] Step S980: Classify and store the sensor data based on the drift type of the sensor data.

[0185] Optionally, for sensor data of three different drift types—no drift, fault drift, and operating condition drift—the corresponding data classification model is matched for classification and storage.

[0186] Optionally, for non-drifted sensor data, priority should be given to ensuring routine monitoring needs. For fault-drifted sensor data, fault diagnosis should be prioritized. For sensor data with drifting operating conditions, the data classification model should be updated before data classification.

[0187] Optionally, step S980 includes steps S981 to S983.

[0188] Step S981: When the drift type of the sensor data is no drift, the sensor data is classified and stored using a preset first data classification model.

[0189] Optionally, in the first data classification model, the sensor sub-data corresponding to each component is stored in a storage partition that is simultaneously combined with the actual operating condition parameters and the component itself.

[0190] Furthermore, in the first data classification model, the sensor sub-data corresponding to each component is further classified and stored according to the data type, and index data including component identifier, actual operating condition parameter combination, data type and timestamp is generated.

[0191] Step S982: When the drift type of the sensor data is fault drift, the sensor data is classified and stored using a preset second data classification model.

[0192] The second data classification model is used to identify the fault types of sensor data and classify and store the sensor data based on the fault types.

[0193] For example, the second data classification model can be a pre-trained decision tree classification model, support vector machine classification model, or neural network model, etc.

[0194] Optionally, in the second data classification model, the fault type of the sensor data is identified, and then the sensor sub-data related to the fault type of the sensor data is stored in a preset fault drift data storage partition, and index data including fault type identifier and timestamp is generated.

[0195] Optionally, in the second data classification model, the cause and level of the fault are determined based on the sensor data. Then, the sensor sub-data related to the fault type of the sensor data is stored in the fault drift data storage partition corresponding to the preset fault level, and index data including fault type identifier, fault level, fault component identifier and timestamp is generated.

[0196] Step S983: When the drift type of the sensor data is operating condition drift, update the second data classification model based on the sensor data, and use the updated first data classification model to classify and store the sensor data.

[0197] Because operating condition drift is continuous, the second data classification model can be updated multiple times.

[0198] Optionally, the second data classification model can be a pre-trained decision tree classification model, support vector machine classification model, or neural network model, etc.

[0199] Optionally, the current sensing data can be identified as sensing data with operating condition drift, and the second data classification model can be retrained using the current sensing data to update the second data classification model.

[0200] By using the aforementioned differentiated classification and storage method, we can avoid the second data classification model from misidentifying the fault type and improve the rationality and orderliness of data classification and storage, which facilitates subsequent data retrieval, analysis and equipment operation and maintenance.

[0201] Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. For example... Figure 2 As shown, the electronic device 400 includes: one or more processors 410 and a memory 420. Figure 2 Take a processor 410 as an example.

[0202] Alternatively, the processor 410 and the memory 420 can be connected via a bus or other means. Figure 2 Taking the example of a connection between China and Israel via a bus.

[0203] Optionally, the processor 410 is configured to determine a task type and a task feature vector based on the current data processing task, wherein the task type is uploading data or storing data; acquire the data to be processed, space environment monitoring information, and resource constraint information corresponding to the data processing task based on the task type; determine a data feature vector based on the data to be processed, and determine a space environment feature vector based on the space environment monitoring information based on the task type; determine a resource constraint feature vector based on the resource constraint information based on the task constraint information; determine the constraint information corresponding to the data processing task based on the task feature vector, data feature vector, space environment feature vector, and resource constraint feature vector; determine the encoding granularity of the data to be processed based on the constraint information; encode the data to be processed based on the encoding granularity and constraint information to obtain encoded data, and process the encoded data based on the task type.

[0204] Optionally, the memory 420, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules of the adaptive data encoding method in the embodiments of this application. The processor 410 executes various functional applications and data processing of the electronic device 400 by running the non-volatile software programs, instructions, and modules stored in the memory 420, thereby implementing the adaptive data encoding method of the above-described method embodiments.

[0205] Optionally, the memory 420 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device 400, etc. Furthermore, the memory 420 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. Optionally, the memory 420 may include memory remotely located relative to the processor 410, and these remote memories may be connected to the controller via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0206] Optionally, one or more modules are stored in memory 420, and when executed by one or more processors 410, they perform the adaptive data encoding method in any of the above method embodiments, for example, the method described above. Figure 1 The method steps S100 to S800.

[0207] Please refer to Figure 3 , Figure 3This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. The computer-readable storage medium 500 stores program code 510, which can be called by a processor to execute the adaptive data encoding method described in the above method embodiments.

[0208] The computer-readable storage medium 500 may be an electronic storage device such as flash memory, electrically erasable programmable read-only memory (EEPROM), hard disk, or read-only memory (ROM). Optionally, the computer-readable storage medium includes a non-volatile computer-readable medium. The computer-readable storage medium 500 has storage space for program code that performs any of the method steps of the adaptive data encoding method described above. This program code can be read from or written to one or more computer program products. The program code may, for example, be compressed in a suitable form.

[0209] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described adaptive data encoding method.

[0210] In summary, this application provides an adaptive data encoding method, electronic device, and storage medium. The adaptive data encoding method includes: determining a task type and a task feature vector based on the current data processing task, where the task type is uploading data or storing data; acquiring the data to be processed, space environment monitoring information, and resource constraint information corresponding to the data processing task based on the task type; determining a data feature vector based on the data to be processed and a space environment feature vector based on the space environment monitoring information based on the task type; determining a resource constraint feature vector based on the resource constraint information based on the task type; determining the constraint information corresponding to the data processing task based on the task feature vector, data feature vector, space environment feature vector, and resource constraint feature vector based on the task type; determining the encoding granularity of the data to be processed based on the constraint information; encoding the data to be processed based on the encoding granularity and constraint information to obtain encoded data; and processing the encoded data based on the task type. This application, by determining the encoding granularity of the data to be processed based on constraint information, enables adaptive data encoding according to the space environment, improving the flexibility of data encoding and its adaptability to extreme space environments.

[0211] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An adaptive data encoding method, characterized in that, include: The task type and task feature vector are determined based on the current data processing task, wherein the task type is uploading data or storing data; Acquire the data to be processed, space environment monitoring information, and resource constraint information corresponding to the data processing task; Determine the data feature vector based on the data to be processed. Determine the space environment feature vector based on the aforementioned space environment monitoring information; Determine the resource constraint feature vector based on the aforementioned resource constraint information; The constraint information corresponding to the data processing task is determined based on the task feature vector, the data feature vector, the spatial environment feature vector, and the resource constraint feature vector. The encoding granularity of the data to be processed is determined based on the constraint information; The data to be processed is encoded based on the encoding granularity and the constraint information to obtain encoded data, and the encoded data is processed based on the task type.

2. The adaptive data encoding method according to claim 1, characterized in that, The task feature vector includes at least one feature parameter selected from task security level, task real-time performance level, and task importance level; and / or The data feature vector includes at least one of the following feature parameters: data type parameter, access frequency, lifespan, and security level; and / or The spatial environment feature vector includes at least one feature parameter selected from equipment temperature, radiation intensity, channel bit error rate, and link signal-to-noise ratio; and / or The resource constraint feature vector includes at least one of the following feature parameters: processor computing power margin, storage capacity margin, power supply margin, and available bandwidth.

3. The adaptive data encoding method according to claim 2, characterized in that, Determining the encoding granularity of the data to be processed based on the constraint information includes: The initial coding granularity is determined based on the data feature vector in the constraint information and multiple preset coding granularities; The environmental granularity adjustment amount is determined based on the aforementioned spatial environment feature vector; The task granularity adjustment amount is determined based on the task feature vector; The resource constraint adjustment amount is determined based on the aforementioned resource constraint feature vector; The final coding granularity is determined based on the initial coding granularity, the environment granularity adjustment amount, the task granularity adjustment amount, and the resource constraint adjustment amount.

4. The adaptive data encoding method according to claim 3, characterized in that, The space environment feature vector includes equipment temperature, radiation intensity, and single-particle flip rate. Determining the environmental granularity adjustment based on the space environment feature vector includes: The environmental particle size adjustment amount is determined based on the ratio of the device temperature to a preset temperature threshold and the ratio of the radiation intensity to a preset radiation intensity threshold.

5. The adaptive data encoding method according to claim 3, characterized in that, The task feature vector includes the task real-time level, and the step of determining the task granularity adjustment amount based on the task feature vector includes: The task granularity adjustment amount is determined based on the task real-time level in the task feature vector, wherein the task granularity adjustment amount is negatively correlated with the task real-time level.

6. The adaptive data encoding method according to claim 3, characterized in that, The resource constraint feature vector includes processor computing power margin, and determining the resource constraint adjustment amount based on the resource constraint feature vector includes: The resource constraint adjustment amount is determined based on the processor computing power margin in the resource constraint feature vector, wherein the resource constraint adjustment amount is positively correlated with the processor computing power margin.

7. The adaptive data encoding method according to claim 1, characterized in that, The process of encoding the data to be processed based on the encoding granularity and the constraint information to obtain encoded data, and then processing the encoded data based on the task type, includes: Based on the encoding granularity, the data to be processed is divided into multiple data blocks to be processed; The target encoding method for the data block to be processed is determined based on the encoding granularity and the constraint information. Each of the data blocks to be processed is encoded using the target encoding method to obtain multiple encoded data blocks; The encoded data is obtained by concatenating multiple encoded data blocks, and the encoded data is processed based on the task type.

8. The adaptive data encoding method according to claim 7, characterized in that, Determining the target encoding method for the data block to be processed based on the encoding granularity and the constraint information includes: Determine the constraint feature vector based on the constraint information; A performance quantization feature vector is determined for each of the preset multiple encoding methods corresponding to the encoding granularity. The performance quantization feature vector includes compression ratio, encryption strength, error correction capability, computing power consumption, and encoding latency. Based on the constraint feature vector and the performance quantization feature vector, one of the multiple encoding methods is selected as the target encoding method.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the adaptive data encoding method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an executable program, which is executed by a processor to implement the adaptive data encoding method as described in any one of claims 1 to 8.

Images