Airborne test system sensor configuration evaluation method and system based on service matching and matrix decomposition

Abstract

The invention provides an airborne test system sensor configuration evaluation method and system. The method comprises the following steps: establishing a test task database which comprises a plurality of test task sub-libraries; establishing a sensor database, wherein the sensor database comprises a plurality of sensor sub-libraries corresponding to the test task sub-libraries; for a target test task in the test task sub-library, screening a sensor from the corresponding sensor sub-library to generate an alternative sensor group; establishing a use record matrix, a random initialization task feature matrix and a sensor feature matrix based on a historical configuration instance; constructing a partial order pair set; defining an optimization target; iteratively updating the task feature matrix and the sensor feature matrix based on the partial order pair set until convergence; and calculating a scoring matrix based on the converged task feature matrix and the converged sensor feature matrix, and selecting the sensor with the highest score from the alternative sensor group as the sensor of the target test task according to the calculated scoring matrix.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of airborne test system configuration, and in particular to a sensor selection evaluation method and system for airborne test system based on business matching and matrix decomposition. BACKGROUND

[0002] An airborne test system is widely used in the fields of aircraft performance evaluation, flight test verification and fault diagnosis, and is a core support system in the aircraft research and development, flight test and operation stages. The core task of the airborne test system is to acquire flight data through sensors to support system state monitoring and key performance analysis. However, the configuration process of the airborne test system usually involves a large number of sensor types, and how to optimize the selection of sensors while meeting the test requirements to improve the accuracy of data acquisition and the efficiency of the system is an important problem in current research.

[0003] At present, the selection of sensors in flight test tasks mainly relies on expert experience, which has derived a series of problems: the configuration scheme is highly subjective, the calculation process is complex, and it is difficult to meet the diversified needs of different tasks. Traditional intelligent configuration methods mostly use system analysis methods, which greatly depend on the subjective judgment and experience of experts in constructing hierarchical structure, determining index weight and carrying out pairwise comparison and judgment. The judgment results given by different experts may be different, resulting in obvious subjectivity of the final conclusion. Even through consistency test, it is also difficult to completely eliminate the deviation caused by subjective factors. With the continuous improvement of the intelligence level of airborne systems, many sensor configuration examples have been accumulated. However, data-driven methods represented by neural networks need a large amount of data to update neuron weights, and have very high requirements for data granularity and cleanliness, and are not suitable for the sensor configuration selection scene of airborne test systems.

[0004] Therefore, in practice, there is an urgent need for a sensor configuration evaluation method that can meet business demand matching and realize data-driven optimization, so as to improve the configuration efficiency. SUMMARY

[0005] The summary is provided to introduce some concepts in a simplified form that will be further described in the following detailed description. The summary is not intended to identify key or essential features of the claimed subject matter nor is it intended to be used to determine the scope of the claimed subject matter.

[0006] As described above, the airborne test system collects flight data (such as temperature, pressure, vibration, strain, etc.) by sensors for aircraft performance evaluation, fault diagnosis, etc. However, when selecting sensors, previous methods mainly rely on expert experience, which is highly subjective and difficult to meet different task requirements; traditional intelligent methods also rely on expert judgment, which is prone to bias; and data-driven methods such as neural networks require a large amount of data and have very high requirements for data granularity and cleanliness, which are not suitable for sensor configuration selection scenarios of airborne test systems.

[0007] The present application proposes an innovative airborne test system sensor selection evaluation method, which combines business matching and matrix decomposition technology to comprehensively improve the scientificity and intelligent level of configuration decision-making. The method first accurately connects the sensor configuration and the requirements of the specific test task based on the business matching mechanism, ensuring its adaptability and pertinence; then, the matrix decomposition technology is introduced to analyze the historical sensor actual selection cases, extract the configuration preferences under different tasks, and mine the potential selection patterns. This method is based on case data-driven, which not only effectively reduces the dependence on expert experience and reduces the cost of manual evaluation, but also provides solid data support for performance evaluation and fault diagnosis during flight testing, promoting the development of airborne test systems towards high efficiency and intelligence.

[0008] According to an embodiment of the present application, an airborne test system sensor configuration evaluation method is provided, comprising: establishing a test task database, the test task database comprising a plurality of test task sub-bases, each test task sub-base comprising one or more test tasks, and each test task comprising one or more test requirements; establishing a sensor database, the sensor database comprising a plurality of sensor sub-bases corresponding to the test task sub-bases, each sensor sub-base comprising one or more sensors, and each sensor comprising one or more sensor attributes; for each target test task in the test task sub-bases, filtering all sensors that meet the test requirements of the target test task from the sensor sub-base corresponding to the test task sub-base to generate a group of candidate sensors; based on historical configuration instances, establishing a usage record matrix, and randomly initializing a task feature matrix and a sensor feature matrix; based on the usage record matrix, constructing a set of partial order pairs, wherein a partial order pair is used to represent the priority relationship of different sensors under the same test task; defining an optimization target for solving the task feature matrix and the sensor feature matrix; based on the set of partial order pairs, iteratively updating the task feature matrix and the sensor feature matrix until convergence; and based on the converged task feature matrix and the converged sensor feature matrix, calculating a score matrix, and for each target test task, selecting the sensor with the highest score from the candidate sensor group as the sensor of the target test task according to the calculated score matrix.

[0009] According to another embodiment of the present application, there is provided a system for airborne test system sensor configuration evaluation, comprising: a mission database establishing module configured to establish a test mission database, the test mission database comprising a plurality of test mission sub-libraries, each test mission sub-library comprising one or more test missions, each test mission comprising one or more test requirements; a sensor database establishing module configured to establish a sensor database, the sensor database comprising a plurality of sensor sub-libraries corresponding to the test mission sub-libraries, each sensor sub-library comprising one or more sensors, each sensor comprising one or more sensor attributes; an alternative sensor group generating module configured to, for each of target test missions in the test mission sub-libraries, screen sensors from the sensor sub-library corresponding to the test mission sub-library that satisfy all test requirements of the target test mission, and generate an alternative sensor group; a matrix establishing module configured to establish a usage record matrix based on historical configuration instances, and randomly initialize a mission feature matrix and a sensor feature matrix; a partial order pair set constructing module configured to construct a partial order pair set based on the usage record matrix, wherein a partial order pair is used to represent a priority relationship between different sensors under a same test mission; an optimization objective defining module configured to define an optimization objective for solving the mission feature matrix and the sensor feature matrix; a matrix convergence module configured to iteratively update the mission feature matrix and the sensor feature matrix based on the partial order pair set until convergence; and a sensor selecting module configured to calculate a score matrix based on the converged mission feature matrix and the converged sensor feature matrix, and for each target test mission, select a sensor with the highest score from the alternative sensor group as the sensor for the target test mission according to the calculated score matrix.

[0010] According to another embodiment of the present application, there is provided a mobile device, comprising: a memory; and a system for airborne test system sensor configuration evaluation as described above.

[0011] These and other features and advantages will become apparent to those of ordinary skill in the art upon reading the following detailed description, taken in conjunction with the drawings in which: BRIEF DESCRIPTION OF DRAWINGS

[0012] For a more complete understanding of the foregoing features of the present application, reference is made to the detailed description taken in conjunction with the accompanying drawings in which certain typical aspects of the application are illustrated. However, it is to be noted that the drawings merely provide typical aspects of the application and should not be construed to limit the scope of the application, as the description can admit to other equally effective aspects.

[0013] Figure 1 A flowchart of a method 100 for evaluating sensor configuration for an on-board test system, according to one embodiment of the present application, is shown.

[0014] Figure 2 A block diagram of a system 200 for evaluating sensor configuration for an on-board test system, according to one embodiment of the present application, is shown.

[0015] Figure 3 A block diagram 300 of an exemplary computing device, according to one embodiment of the present application, is shown. DETAILED DESCRIPTION

[0016] The present application will be described in detail below with reference to the attached drawings.

[0017] Figure 1 A flowchart of a method 100 for evaluating sensor configuration for an on-board test system, according to one embodiment of the present application, is shown.

[0018] Generally, the present application 100 includes the following aspects:

[0019] Database establishment: test task database (such as temperature, pressure test task, etc., each task has measurement range, accuracy, etc. requirements) and sensor database (corresponding to the task type has different sensors, each sensor has model, measurement range, etc. attributes) are established respectively.

[0020] Screening of alternative sensors: according to the requirements of the test task, the sensors that meet all the requirements are selected from the corresponding sensor library to form an alternative group.

[0021] Establishment of correlation matrix: a use record matrix (recording the number of times of using each sensor for each task) is established according to the historical use record, and then the task feature matrix and the sensor feature matrix are randomly initialized.

[0022] Construction of partial order pair: from the use record matrix, the relationship such as "A sensor is used more than B sensor for a certain task" is found to form a partial order pair.

[0023] Optimization matrix: setting the optimization goal, updating the task feature matrix and the sensor feature matrix by gradient ascent method until stable.

[0024] Select the optimal sensor: according to the score matrix calculated based on the optimized matrix, select the highest score from the alternative sensors.

[0025] The following specifically describes each step of the method 100.

[0026] At 101, a test task database is established. Generally speaking, the test task database is used to store the basic information and requirement parameters of various test tasks, and provides clear business benchmarks for subsequent sensor screening. Specifically, the test task database includes multiple test task sub-libraries, each of which corresponds to a type of test task. In addition, each test task sub-library includes one or more test tasks, and each test task includes one or more test requirements (such as measurement range, accuracy, frequency, installation position, etc.).

[0027] According to an embodiment of the present application, a test task database D Tasks is established.

[0028] Wherein D Tasks includes various test task sub-libraries T, such as temperature test task sub-libraries, pressure test task sub-libraries, vibration test task sub-libraries, strain test task sub-libraries, etc., denoted as

[0029] Wherein the test task sub-library T includes a series of test tasks, denoted as Wherein N T is the number of test tasks in the test task sub-library T;

[0030] Wherein the test task t includes a series of test requirements, including test parameter name, measurement unit, measurement range upper and lower limits, measurement accuracy, test frequency, measurement point installation position, measurement point environment temperature range upper and lower limits, measurement point installation size limit, sensor shape form limit, etc.

[0031] At 102, a sensor database is established. Generally speaking, the sensor database is used to store the performance parameters and attribute information of various sensors, and forms a corresponding relationship with the test task database, providing a data basis for "business requirement-sensor performance" matching. Specifically, the sensor database includes multiple sensor sub-libraries, which correspond one by one to the test task sub-libraries (such as temperature sensor sub-libraries corresponding to temperature test task sub-libraries); each sensor sub-library includes one or more sensors, and each sensor includes sensor attributes (such as sensor model, measurement unit, measurement range, measurement accuracy, frequency response, use temperature range, volume size, shape characteristics, power supply requirement, sensor price, and sensor quality expert rating, etc.).

[0032] According to an embodiment of the present application, a sensor database D Sensors is established.

[0033] wherein D Sensors Each sensor sub-library S corresponds to a test task sub-library, such as a temperature sensor sub-library, a pressure sensor sub-library, a vibration sensor sub-library, a strain sensor sub-library, and the like, denoted as

[0034] The sensor sub-library S includes a series of sensors, denoted as wherein N S is the number of sensors in the sensor sub-library S;

[0035] The sensor s includes a series of sensor attributes, including sensor model, measurement unit, measurement range, measurement accuracy, frequency response, temperature range, volume size, shape characteristics, power supply requirements, sensor price, and sensor quality expert rating.

[0036] At 103, for each target test task in the target test task sub-library, all sensors that meet the test requirements of the target test task are screened from the sensor sub-library corresponding to the test task sub-library to generate a candidate sensor group. Generally, step 103 screens the sensors that meet the test requirements from the sensor database through precise matching of "test task requirements-sensor performance" to form a candidate sensor group, providing a candidate set for subsequent optimization.

[0037] According to one embodiment of the present application, for a test task t i in the test task sub-library T, there are several test requirements, and sensors that meet the test requirements are screened from the sensor sub-library s corresponding to the test task sub-library according to each test requirement, and the sensors that meet all the test requirements are added to the candidate sensor group to finally form the candidate sensor group.

[0038] At 104, a usage record matrix is established based on historical configuration instances, and a task feature matrix and a sensor feature matrix are randomly initialized, wherein the usage record matrix records the historical usage times of each target test task on the sensors in the candidate sensor group, the task feature matrix is used to represent the test requirements of each target test task, and the sensor feature matrix is used to represent the sensor attributes of the sensors in the candidate sensor group.

[0039] As can be seen, the usage record matrix converts the historical sensor configuration instances (usage times of tasks on sensors) into a calculable matrix form, providing a basis for subsequent mining of configuration rules. The task feature matrix and the sensor feature matrix convert the high-dimensional and complex "task-sensor" relationship into low-dimensional and interpretable feature associations by extracting "potential demand features of test tasks" and "potential capability features of sensors", realizing the mining of configuration rules.

[0040] According to one embodiment of the present invention, a usage record matrix is ​​established based on historical configuration instances. Simultaneously, the task feature matrix is ​​randomly initialized. Sensor feature matrix

[0041] Using a record matrix The element in the i-th row and j-th column of this matrix is ​​denoted as m. i，j , indicating test task t i Using sensors j The number of times;

[0042] Task feature matrix Where t i =[t i1 , t i2 …t iK [ ] represents the task feature vector, initialized randomly;

[0043] Sensor feature matrix Where s j =[s j1 s j2 …s jK [] represents the sensor feature vector, initialized randomly;

[0044] Where K is the number of latent variable dimensions, and K << N T and N S .

[0045] In 105, a set of partial order pairs is constructed based on the use of a record matrix, where the partial order pairs are used to characterize the priority relationship between different sensors under the same test task.

[0046] According to one embodiment of the present invention, based on the use of the record matrix Construct about task t i Partial order pairs (t) i s p s q ), indicating that for test task t i Sensors p Priority is higher than that of sensor s q (That is, for test task t) i Sensors p The number of times the sensor has been used in history is greater than that of the sensor. q (historical usage count), abbreviated as The set of all partially ordered pairs is denoted as in

[0047] The construction method is as follows: initial Row traversal matrix For i-th row, if there are m i,p >m i,q , the triplet (t i , s p , s q ) is added to the set of triplets

[0048] At 106, an optimization objective is defined for solving the task feature matrix and the sensor feature matrix, the optimization objective including a partial order relation fitting sub-objective and a regularization sub-objective. The partial order relation fitting sub-objective is used to correct the deviation caused by the randomness of the initial feature matrix, and the regularization sub-objective is used to avoid overfitting of the historical data by the feature matrix and ensure the generalization ability of the evaluation.

[0049] In practice, the optimization objective is a parameter adjustment criterion for the task feature matrix and the sensor feature matrix, and the core is to constrain through a mathematical model, so that the score result generated by the feature matrix is more in line with the historical configuration rule, while avoiding overfitting. In other words, the optimization objective is used to guide the iterative update of the task feature matrix and the sensor feature matrix, so that the generated score matrix can accurately reflect the priority relationship in the partial order pair, while ensuring the stability of the feature matrix.

[0050] According to an embodiment of the present application, an optimization objective is defined for solving the task feature matrix and the sensor feature matrix The optimization objective is:

[0051]

[0052] Wherein, for the optimization sub-objective P(> t |T, S) (i.e., the partial order relation fitting sub-objective), there is:

[0053]

[0054] Wherein, and σ(x) = 1 / (1+e -x ), where k∈(0, 1] is an expert coefficient, indicating the degree of recognition of the expert for the ranking, and if there is no expert coefficient, the default value is 0.8;

[0055] Wherein, for the optimization sub-objective P(T, S) (i.e., the regularization sub-objective), it is assumed that the probability prior of the parameters of the task feature matrix and the sensor feature matrix obeys a Gaussian distribution, and there is:

[0056] lnP(T, S) = λ||T, S|| 2

[0057] Wherein, λ is a regularization coefficient.

[0058] At 107, iteratively update the task feature matrix and the sensor feature matrix by a gradient ascent method, wherein the gradient in the gradient ascent method is based on the partial order pair set.

[0059] Specifically, the gradient ascent method takes the partial order relationship fitting sub-object (fitting the historical configuration priority) and the regularization sub-object (avoiding overfitting) as the optimization direction, and by continuously adjusting the parameters of the task feature matrix and the sensor feature matrix, the final generated score matrix gradually conforms to the sensor priority relationship in the historical use records. When the iteration converges, the parameters of the feature matrix no longer change significantly, which means that the score result generated by the feature matrix has stably reflected the actual adaptation rule of the "test task-sensor", and the deviation caused by random initialization is eliminated.

[0060] According to an embodiment of the present application, the entire partial order pair set is used in a cycle Calculate the gradient, and update the parameters of the matrix T and S based on the gradient ascent method:

[0061]

[0062] Until the task feature matrix The sensor feature matrix Converges.

[0063] Wherein, a is the learning rate.

[0064] For steps 106 and 107, specifically, in the above step 104, the task feature matrix (characterizing the potential requirements of the test task) and the sensor feature matrix (characterizing the potential capabilities of the sensor) are "randomly initialized" to generate. This randomness causes the initial matrix to not accurately reflect the real task requirement features and sensor capability features. By optimization (steps 106 and 107), the element values of the task feature matrix and the sensor feature matrix can be adjusted, so that the final generated score matrix not only conforms to the partial order pair (historical rule), but also can mine the potential adaptation mode.

[0065] At 108, calculate the score matrix based on the converged task feature matrix and the converged sensor feature matrix, and for each target test task, select the sensor with the highest score from the candidate sensor group as the sensor of the target test task according to the calculated score matrix. Wherein, the score matrix is the "associated product" of the task feature matrix and the sensor feature matrix, which is used to quantify the adaptation degree of each sensor to each test task, and provides a direct basis for configuration evaluation. That is, the score matrix is used to represent the adaptability of the sensors in the candidate sensor group to the test task.

[0066] According to an embodiment of the present application, according to the optimized task feature matrix and sensor feature matrix computing a score matrix selecting the sensor with the highest score from the candidate sensor set as the sensor for the test task according to the score matrix.

[0067] score matrix There are:

[0068]

[0069] score matrix A total order relation is defined with respect to test task t, denoted as t The element in the i-th row and j-th column of the matrix is denoted as represents the score of sensor s i for test task t j ;

[0070] wherein,

[0071] For test task t i , take The i-th row has a list The larger the element in the list indicates that the sensor s i corresponding to the element is more recommended for test task t

[0072] Figure 2 A block diagram of a system 200 for sensor configuration evaluation of an onboard test system according to one embodiment of the present application is shown.

[0073] As shown, the system 200 can include a test task database establishing module 201, a sensor database establishing module 202, a candidate sensor set generating module 203, a matrix establishing module 204, a partial order pair set constructing module 205, an optimization objective defining module 206, a matrix converging module 207, and a sensor selecting module 208. It is well understood by those skilled in the art that the division of the above modules is only for the purpose of illustration. The functions of one or more of the above modules can be combined into a single module or split into more modules. Also, one or more of the above modules can be implemented in software, hardware, or a combination thereof. Furthermore, the data flow between the modules can be in a manner known in the art, which is not discussed herein.

[0074] According to one embodiment of the present application, the test task database establishing module 201 is configured to establish a test task database. The test task database contains a plurality of test task sub-libraries, each of which corresponds to a type of test task. In addition, each test task sub-library contains one or more test tasks, each of which contains one or more test requirements.

[0075] According to an embodiment of the present application, the sensor database establishing module 202 is configured to establish a sensor database. The sensor database comprises a plurality of sensor sub-libraries, each of which corresponds to a test task sub-library; each sensor sub-library comprises one or more sensors, each of which comprises sensor attributes.

[0076] According to an embodiment of the present application, the alternative sensor group generating module 203 is configured to, for each of the target test tasks in the test task sub-library, screen sensors from the sensor sub-library corresponding to the test task sub-library that satisfy all test requirements of the target test task, and generate an alternative sensor group.

[0077] According to an embodiment of the present application, the matrix establishing module 204 is configured to establish a usage record matrix, a task feature matrix and a sensor feature matrix based on historical configuration instances, wherein the usage record matrix records the historical usage times of the sensors in the alternative sensor group by each target test task, the task feature matrix is used to characterize the test requirements of each target test task, and the sensor feature matrix is used to characterize the sensor attributes of the sensors in the alternative sensor group.

[0078] According to an embodiment of the present application, the partial order pair set constructing module 205 is configured to construct a partial order pair set based on the usage record matrix, wherein the partial order pair is used to characterize the priority relationship of different sensors under the same test task.

[0079] According to an embodiment of the present application, the optimization target defining module 206 is configured to define an optimization target for solving the task feature matrix and the sensor feature matrix, the optimization target comprising a partial order relationship fitting sub-target and a regularization sub-target. The partial order relationship fitting sub-target is used to correct the deviation caused by the randomness of the initial feature matrix, and the regularization sub-target is used to avoid overfitting of the feature matrix to the historical data and ensure the generalization ability of the evaluation.

[0080] According to an embodiment of the present application, the matrix convergence module 207 is configured to iteratively update the task feature matrix and the sensor feature matrix until convergence by a gradient ascent method, wherein the gradient in the gradient ascent method is based on the partial order pair set.

[0081] According to an embodiment of the present application, the sensor selecting module 208 is configured to calculate a score matrix based on the converged task feature matrix and the converged sensor feature matrix, and for each target test task, select the sensor with the highest score from the alternative sensor group as the sensor of the target test task according to the calculated score matrix. The score matrix is used to characterize the adaptability of the sensors in the alternative sensor group to the test task. Specific embodiments:

[0083] The following example illustrates the application of this invention in pressure sensor selection during flight test missions, combined with... Figure 1 The embodiments of the present invention are described below.

[0084] Step 1: Establish the test task database D Tasks ;

[0085] Establish a stress testing task sub-library T pressure It contains a series of stress test tasks t, denoted as in For the stress testing task sub-library T pressure The number of test tasks in the system;

[0086] The pressure test task includes a series of test requirements, including measurement range requirements, measurement accuracy requirements, measurement sensitivity requirements, response time requirements, test frequency requirements, weight requirements, measurement point installation location requirements, upper and lower limits of the measurement point ambient temperature range, measurement point installation size limitations, and sensor shape limitations.

[0087] Step 2: Establish sensor library D Sensors ;

[0088] Establish a pressure sensor sub-library S pressure It contains a series of sensors, denoted as in For sensor sub-library S pressure The number of sensors in the system;

[0089] The sensor s includes a series of sensor attributes, including measurement range, measurement accuracy, measurement sensitivity, response time, test frequency, weight, upper and lower limits of the ambient temperature range of the measurement point, installation dimensions of the measurement point, and sensor shape.

[0090] Step 3: Select candidate pressure sensor groups based on the requirements of the pressure test task;

[0091] Test Task Sublibrary T pressure The stress test task t in i It contains several test requirements, and each test requirement is selected from the pressure sensor sub-library S. pressure Sensors that meet the test requirements are selected, and those that meet all the test requirements are added to the candidate pressure sensor group, thus forming the final candidate pressure sensor group.

[0092] Step 4: Create a record matrix Randomly initialize the task feature matrix Randomly initialize the sensor feature matrix

[0093] Step 5: Based on the record matrix used Construct partial order pair (t i , s i ) for stress test task t p , s q , which means that for test task t i , the priority of sensor s p is higher than that of sensor s q , which is denoted as The whole set of partial order pairs is denoted as Wherein

[0094] Step 6: Define the optimization target for solving the stress task feature matrix and the stress sensor feature matrix The optimization target is:

[0095]

[0096] Step 7: Use the whole set of partial order pairs in a loop Calculate the gradient, update the parameters of the matrix T, S based on the gradient ascent method:

[0097]

[0098] Until the stress task feature matrix and the stress sensor feature matrix converge.

[0099] Step 8: According to the optimized task feature matrix and the sensor feature matrix Calculate the T pressure -S pressure score matrix Wherein According to the score matrix, select the sensor with the highest score in the alternative stress sensor group as the sensor for the test task.

[0100] Figure 3 A block diagram 300 of an exemplary computing device according to an embodiment of the present application is shown, which is one example of a hardware device applicable to aspects of the present application. Reference is made to Figure 3A computing device 300 will now be described, which is one example of a hardware device that can be applied to aspects of the present application. The computing device 300 can be any machine that is capable of implementing processing and / or computation, which can be, but is not limited to, a workstation, a server, a desktop computer, a laptop computer, a tablet computer, a personal digital assistant, a smartphone, an in-vehicle computer, or any combination thereof. The computing device 300 can include components that can be connected or communicate via one or more interfaces and buses 302. For example, the computing device 300 can include a bus 302, one or more processors 304, one or more input devices 306, and one or more output devices 308. The one or more processors 304 can be any type of processor and can include, but are not limited to, one or more general-purpose processors and / or one or more special-purpose processors (e.g., specialized processing chips). The input devices 306 can be any type of device capable of inputting information to the computing device and can include, but are not limited to, a mouse, a keyboard, a touchscreen, a microphone, and / or a remote controller. The output devices 308 can be any type of device capable of presenting information and can include, but are not limited to, a display, a speaker, a video / audio output terminal, a vibrator, and / or a printer. The computing device 300 can also include or be connected to a non-transitory storage device 310, which can be any storage device that is non-transitory and capable of implementing data storage, and which can include, but is not limited to, a disk drive, an optical storage device, a solid-state memory, a floppy disk, a flexible disk, a hard disk, a magnetic tape or any other magnetic medium, an optical disk or any other optical medium, a ROM (read only memory), a RAM (random access memory), a cache memory and / or any other memory chip or cartridge, and / or any other medium from which the computer can read data, instructions, and / or code. The non-transitory storage device 310 can be detached from the interface. The non-transitory storage device 310 can have data / instructions / code for implementing the above-described methods and steps. The computing device 300 can also include a communication device 312. The communication device 312 can be any type of device or system capable of implementing communication with internal devices and / or communication with a network, and can include, but is not limited to, a modem, a network card, an infrared communication device, a wireless communication device, and / or a chipset such as a Bluetooth device, an IEEE 802.11 device, a WiFi device, a WiMax device, a cellular communication device, and / or the like.

[0101] The bus 302 can include, but is not limited to, an Industry Standard Architecture (ISA) bus, a MicroChannel Architecture (MCA) bus, an Enhanced ISA (EISA) bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0102] The computing device 300 can also include a working memory 314, which can be any type of memory capable of storing instructions and / or data used to operate the processor 304 and can include, but is not limited to, random access memory (RAM), and / or read-only memory (ROM).

[0103] Software components can reside in the working memory 314, including, but not limited to, an operating system 316, one or more applications 318, drivers, and / or other data and code. Instructions implementing the methods and steps of the present application can be contained in the one or more applications 318, and the instructions of the one or more applications 318 can be read and executed by the processor 304 to implement the methods 300 of the present application.

[0104] It should also be recognized that variations can be made in light of specific needs. For example, custom hardware might also be used, and / or specific components can be implemented in hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. Additionally, connections with other computing devices, such as network input / output devices, can be employed. For example, portions or all of the disclosed methods and devices can be implemented using logic and algorithms according to the present application by programming hardware having an assembly language or a hardware programming language (e.g., VERILOG, VHDL, C++).

[0105] While aspects of the present application have been described with reference to the attached figures, it is to be understood that the methods and devices are merely examples and that the scope of the present application is not limited to these aspects, but rather only to the claims and their equivalents. Various components can be omitted or substituted for equal components. Additionally, the steps can be implemented in an order different from that described. Furthermore, the various components can be combined in various ways. It is also important to note that many of the components described can be implemented using equivalents.

Claims

1. A method for evaluating sensor configuration of an airborne test system, comprising: establishing a test task database comprising a plurality of test task sub-libraries, each test task sub-library comprising one or more test tasks, each test task comprising one or more test requirements; establishing a sensor database comprising a plurality of sensor sub-libraries corresponding to the plurality of test task sub-libraries, each sensor sub-library comprising one or more sensors, each sensor comprising one or more sensor attributes; for each of target test tasks in a test task sub-library, screening sensors from a sensor sub-library corresponding to the test task sub-library that satisfy all test requirements of the target test task, to generate a candidate sensor group; establishing a usage record matrix based on historical configuration instances, and randomly initializing a task feature matrix and a sensor feature matrix; constructing a set of partial order pairs based on the usage record matrix, wherein a partial order pair is used to represent a priority relationship between different sensors under a same test task; defining an optimization objective for solving the task feature matrix and the sensor feature matrix; iteratively updating the task feature matrix and the sensor feature matrix based on the set of partial order pairs until convergence; and calculating a score matrix based on the converged task feature matrix and the converged sensor feature matrix, and for each target test task, selecting a sensor with the highest score from the candidate sensor group as the sensor for the target test task according to the calculated score matrix.

2. The method of claim 1, wherein, The usage record matrix records a historical usage number of each target test task to the sensors in the candidate sensor group, the task feature matrix is used to represent test requirements of each target test task, the sensor feature matrix is used to represent sensor attributes of the sensors in the candidate sensor group, and the score matrix is used to represent adaptability of the sensors in the candidate sensor group to test tasks.

3. The method of claim 1, wherein, The optimization objective comprises a partial order relationship fitting sub-objective and a regularization sub-objective.

4. The method of claim 3, wherein, The partial order relationship fitting sub-objective is used to correct score deviation caused by random initialization of the task feature matrix and the sensor feature matrix, and the regularization sub-objective is used to avoid overfitting of the task feature matrix and the sensor feature matrix to historical data.

5. The method of claim 1, wherein, The iteratively updating the task feature matrix and the sensor feature matrix until convergence further comprises iteratively updating the task feature matrix and the sensor feature matrix until convergence by a gradient ascent method, wherein a gradient in the gradient ascent method is based on the set of partial order pairs.

6. The method of claim 1, wherein, The test requirements comprise one or more of the following: a test parameter name, a measurement unit, a measurement range upper and lower limit, a measurement precision, a test frequency, a measurement point installation position, a measurement point environment temperature range upper and lower limit, a measurement point installation size limit, and a sensor shape form limit.

7. A system for evaluating sensor configuration of an airborne test system, comprising: a task database establishing module configured to establish a test task database, the test task database comprising a plurality of test task sub-libraries, each test task sub-library comprising one or more test tasks, each test task comprising one or more test requirements; a sensor database establishing module configured to establish a sensor database, the sensor database comprising a plurality of sensor sub-libraries corresponding to the plurality of test task sub-libraries, each sensor sub-library comprising one or more sensors, each sensor comprising one or more sensor attributes; an alternative sensor group generating module configured to, for each of target test tasks in a test task sub-library, screen sensors from a sensor sub-library corresponding to the test task sub-library that satisfy all test requirements of the target test task, and generate an alternative sensor group; a matrix establishing module configured to establish a usage record matrix based on historical configuration instances, and randomly initialize a task feature matrix and a sensor feature matrix; a partial order pair set constructing module configured to construct a partial order pair set based on the usage record matrix, wherein a partial order pair is used to represent a priority relationship between different sensors under a same test task; an optimization objective defining module configured to define an optimization objective for solving the task feature matrix and the sensor feature matrix; a matrix convergence module configured to iteratively update the task feature matrix and the sensor feature matrix based on the partial order pair set until convergence; and a sensor selecting module configured to calculate a score matrix based on the converged task feature matrix and the converged sensor feature matrix, and for each target test task, select a sensor with the highest score from the alternative sensor group as the sensor for the target test task according to the calculated score matrix.

8. The system of claim 7, wherein, The usage record matrix records a historical usage number of each target test task on the sensors in the alternative sensor group, the task feature matrix is used to represent test requirements of each target test task, the sensor feature matrix is used to represent sensor attributes of the sensors in the alternative sensor group, and the score matrix is used to represent adaptability of the sensors in the alternative sensor group to test tasks.

9. The system of claim 8, wherein, Iteratively updating the task feature matrix and the sensor feature matrix until convergence further comprises iteratively updating the task feature matrix and the sensor feature matrix until convergence by a gradient ascent method, wherein a gradient in the gradient ascent method is based on the partial order pair set.

10. A computing device comprising: a memory; the system of any one of claims 7-9.

Images