Unmanned aerial vehicle body intelligent and reliable sensing method based on coevolution

Abstract

The invention discloses an unmanned aerial vehicle body intelligent reliable sensing method based on coevolution, and belongs to the technical field of unmanned aerial vehicle safe flight, and the method comprises the steps: processing the information of an observed unmanned aerial vehicle through a clustering analysis algorithm, and obtaining the physical condition of the swing amplitude of the unmanned aerial vehicle; performing flight reliable sensing processing on the physical condition of the swing amplitude of the unmanned aerial vehicle by using a set co-evolution model to obtain a flight attitude reliable sensing parameter corresponding to the physical condition of the swing amplitude of the unmanned aerial vehicle; constructing an edge computing gateway of the physical condition of the swing amplitude of the observed unmanned aerial vehicle by utilizing the physical condition of the swing amplitude of the unmanned aerial vehicle and the reliable sensing parameters of the flight attitude; the observed unmanned aerial vehicle swing amplitude physical condition edge calculation gateway generates an observed unmanned aerial vehicle swing amplitude physical condition boundary threshold value, the observed unmanned aerial vehicle swing amplitude physical condition boundary threshold value is sent to the intelligent data processing platform with the body, and the model is iterated; the flight safety of the unmanned aerial vehicle is ensured, and the hardware cost is saved.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle (UAV) safe flight technology, and in particular to an embodied intelligent and reliable perception method for UAVs based on co-evolution. Background Art

[0002] When a drone operates autonomously in a complex, unknown or unstructured environment, it first needs to perceive the surrounding environment and determine the flyable area before it can perform reasonable flight path planning and motion control. Therefore, accurate perception and identification of the flyable area is the key to safe navigation of drones, and has broad application prospects in the fields of reconnaissance, search, and monitoring.

[0003] The existing technical solutions for sensing the flight area mainly include two categories. One category is to rely on single-machine perception, that is, to rely on the sensor equipment carried by itself to collect information about the surrounding environment, and then process and calculate the information on the on-board computing device, design corresponding algorithms and software, and realize the perception of the surrounding environment, such as identifying objects, obstacles, roads in the environment, tracking dynamic obstacles, and dividing the space for flight. The second category is collaborative perception, that is, not only relying on the sensor equipment carried by itself, but also with the help of the perception equipment arranged in the surrounding environment, such as the perception module set up on the roadside. These two types of sensor equipment collect environmental information from different perspectives, and then fuse and process them in the computing module to obtain the perception and understanding results of the environment.

[0004] Although the existing UAV collaborative perception has expanded the scope of environmental information detection, it requires the installation of sensor equipment at appropriate locations in the environment in advance, and has low flight stability, which is not suitable for unknown environments that are entered for the first time (such environments cannot arrange sensors at appropriate locations in the flight space in advance). The existing method of using UAV takeoff to achieve perception is not suitable for complex crash environments (such as bridges, wire barriers, and tall trees); a ground station computing module is required to control the flight of the UAV, which increases the equipment cost (such as ground stations); UAVs cannot realize crash risk factor analysis and flexible track autonomous planning, and cannot flexibly and accurately handle human or non-human factors in the physical conditions of the UAV swing amplitude. Therefore, it is necessary to propose a UAV embodied intelligent reliable perception method based on co-evolution. Summary of the invention

[0005] In view of this, the present invention provides a UAV embodied intelligent reliable perception method based on co-evolution, which processes the physical conditions of the UAV swing amplitude for flight reliable perception, ensures the flight safety of the UAV, and saves hardware costs.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A UAV embodied intelligent reliable perception method based on co-evolution includes the following steps:

[0008] Step 1: Integrate the observed UAV crash risk factors, process the observed UAV crash risk factors through cluster analysis algorithm, and obtain the physical conditions of the UAV swing amplitude;

[0009] Step 2: Use the set co-evolution model to perform flight reliable perception processing on the physical condition of the UAV swing amplitude to obtain the flight attitude reliable perception parameter corresponding to the physical condition of the UAV swing amplitude, and the expression is:

[0010]

[0011] in, represents the reliable perception parameter of flight attitude, G represents the weight of the crash affected by human factors, U represents the weight of the crash affected by non-human factors, ξ represents the probability of occurrence of each human factor affecting the crash, Te represents the distribution matrix of the types of human factors affecting the crash, ω represents the probability of occurrence of each non-human factor affecting the crash, Ti represents the distribution matrix of the types of non-human factors affecting the crash, K represents the reliable perception coefficient, represents the embodied intelligence observation error for each physical condition, X j , Y j , Z j Respectively represent the flight attitude roll angle, pitch angle, and yaw angle;

[0012] Step 3, using the physical condition of the swing amplitude of the drone and the reliable perception parameters of the flight attitude to build an edge computing gateway for the physical condition of the swing amplitude of the observed drone; the reliable perception parameters of the flight attitude are the embodied intelligent flight reliable perception information integrated with the reliable flight perception of the physical condition of the swing amplitude of the drone through the co-evolution model;

[0013] Step 4: The observed UAV swing amplitude physical condition edge computing gateway generates the observed UAV swing amplitude physical condition boundary threshold, and sends the observed UAV swing amplitude physical condition boundary threshold to the embodied intelligent data processing platform;

[0014] Step 5, according to the command information sent by the embodied intelligent data processing platform, integrate the flight altitude reliable perception information of the boundary threshold of the physical condition of the swing amplitude of the observed UAV, wherein the flight altitude reliable perception information is the flight control reliable perception information input by the operator of the embodied intelligent data processing platform during the command sending process;

[0015] Step 6, respectively processing the flight attitude reliable perception parameters and the flight altitude reliable perception information through a convolutional neural network, adjusting the parameters of the observed UAV swing amplitude physical condition edge computing gateway according to the flight attitude reliable perception parameters mapped to the UAV swing amplitude physical condition at the same altitude and the flight influence weight of the flight altitude reliable perception information, and iterating the co-evolution model using the observed UAV swing amplitude physical condition edge computing gateway after parameter adjustment;

[0016] Step 7, integrating the coefficient matrix corresponding to the co-evolution model after parameter adjustment and the physical condition of the swing amplitude of the UAV;

[0017] Identify the error source of the coefficient matrix and the flight attitude reliable perception parameter, and identify the interference substitution parameter pair in the physical condition of the UAV swing amplitude, wherein the interference substitution parameter pair includes an error interference parameter for identifying the flight attitude reliable perception parameter, and also includes an allowable interference parameter for identifying the coefficient matrix, wherein the error interference parameter corresponds to the allowable interference parameter;

[0018] Using the interference substitution parameter pair to construct interference substitution data, using the interference substitution data to train a set Transformer model to obtain an interference substitution model, wherein the interference substitution model is used to perform interference substitution on the physical condition of the swing amplitude of the drone;

[0019] The interference substitution model is expressed as:

[0020]

[0021] Among them, D s represents interference replacement data, P1…P n represents the interference substitution parameter matrix, R 11 …R nn represents the influence weight of each interference substitution parameter, F1…F n Indicates whether the interference substitution parameter can be replaced by a matrix. The judgment basis is whether the influence weight of each interference substitution parameter exceeds the preset value. When it is greater than the preset value, it is 1, and when it is less than the preset value, it is 0. j represents the Transformer encoding-decoding efficiency, H c Represents the input sequence of Transformer, i represents the model training count, and I represents the preset total number of model training times.

[0022] Furthermore, in step 6,

[0023] The convolutional neural network uses convolution kernels to perform feature extraction on the reliable perception parameters of flight attitude and the reliable perception information of flight altitude, extracts noise-free reliable perception parameters, and uses a fully connected layer to integrate the two-dimensional feature vectors of the noise-free reliable perception parameters into one-dimensional vectors, and each one-dimensional vector is mapped to a flight impact weight.

[0024] Further, in the step 6, iterating the co-evolution model using the edge computing gateway of the observed UAV swing amplitude physical condition after parameter adjustment includes: integrating a plurality of the flight altitude reliable perception information mapped with the observed UAV swing amplitude physical condition;

[0025] The flight influence weights in different climate environments of the flight attitude reliable perception parameters and the flight altitude reliable perception information respectively integrated with the physical conditions of the swing amplitude of the observed UAV;

[0026] If the mean value of the flight impact weight in different climate environments exceeds the set weight safety parameter, it is determined that the flight attitude reliable perception parameter is interfered by environmental factors, and the flight altitude reliable perception information in different climate environments is normalized to obtain standardized flight reliable perception information, and the physical condition edge computing gateway of the swing amplitude of the observed UAV is adjusted with the standardized flight reliable perception information parameters.

[0027] Furthermore, between step 1 and step 2, the following steps are also included:

[0028] The interference substitution model is applied to preprocess the physical condition of the UAV swing amplitude so that the interference in the physical condition of the UAV swing amplitude can be compensated.

[0029] Furthermore, before step 4, the method further includes:

[0030] The observed UAV swing amplitude physical condition edge computing gateway is updated using the set For loop + Update statement. The update process is used to delete the observed UAV swing amplitude physical condition in the observed UAV swing amplitude physical condition edge computing gateway. The observed UAV swing amplitude physical condition includes the UAV swing amplitude physical condition whose error interference parameter is less than the set parameter range.

[0031] Furthermore, it is realized through the UAV embodied intelligent and reliable perception system based on co-evolution, the system includes:

[0032] A cluster analysis unit, used to integrate the observed UAV crash risk factors, process the observed UAV crash risk factors using a cluster analysis algorithm, and obtain the UAV swing amplitude physical conditions;

[0033] The embodied intelligent flight reliable perception unit is used to perform flight reliable perception processing on the physical condition of the UAV swing amplitude by using the set co-evolution model, obtain flight attitude reliable perception parameters corresponding to the physical condition of the UAV swing amplitude, and construct an edge computing gateway for the physical condition of the observed UAV swing amplitude by using the physical condition of the UAV swing amplitude and the flight attitude reliable perception parameters, wherein the flight attitude reliable perception parameters are embodied intelligent flight reliable perception information integrated with flight reliable perception of the physical condition of the UAV swing amplitude by using the co-evolution model;

[0034] The observed UAV swing amplitude physical condition sending unit is used to generate the observed UAV swing amplitude physical condition boundary threshold value by the observed UAV swing amplitude physical condition edge computing gateway, and send the observed UAV swing amplitude physical condition boundary threshold value to the embodied intelligent data processing platform;

[0035] A flight reliability perception unit for boundary threshold information is used to send information according to the instruction of the embodied intelligent data processing platform, and integrate the flight altitude reliability perception information of the boundary threshold of the physical condition of the swing amplitude of the observed UAV, and the flight altitude reliability perception information is the flight control reliability perception information input by the operator of the embodied intelligent data processing platform during the instruction sending process;

[0036] The flight reliable perception model iteration unit is used to process the flight attitude reliable perception parameters and the flight altitude reliable perception information respectively by using a convolutional neural network, adjust the parameters of the observed UAV swing amplitude physical condition edge computing gateway according to the flight attitude reliable perception parameters mapped to the UAV swing amplitude physical condition at the same altitude and the flight influence weight of the flight altitude reliable perception information, and iterate the co-evolution model using the observed UAV swing amplitude physical condition edge computing gateway after parameter adjustment.

[0037] Furthermore, the flight reliability perception model iteration unit includes:

[0038] A flight altitude reliable perception information unit, used to integrate a plurality of said flight altitude reliable perception information mapped with the physical conditions of the swing amplitude of the observed UAV;

[0039] A flight influence weight unit, used to respectively integrate the flight influence weights in different climate environments of the flight attitude reliable perception parameters and the flight altitude reliable perception information mapped with the physical conditions of the swing amplitude of the observed UAV;

[0040] The information normalization unit is used to determine that the flight attitude reliable perception parameter is interfered by environmental factors if the mean value of the flight impact weight in different climate environments exceeds the set weight safety parameter, and normalize the flight altitude reliable perception information in different climate environments to obtain standardized flight reliable perception information, and adjust the physical condition edge computing gateway of the swing amplitude of the observed UAV with the standardized flight reliable perception information parameters.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The present invention proposes a reliable perception method of drone embodied intelligence based on co-evolution. The physical conditions of the drone swing amplitude contained in the risk factors of the observed drone crash are obtained by segmenting and processing the features of the observed drone crash risk factors. The physical conditions of the drone swing amplitude at this time may include a single human or non-human factor. Then, the initial co-evolution model can be applied to the physical conditions of the drone swing amplitude for flight reliable perception, and the flight attitude reliable perception parameters are obtained. The flight attitude reliable perception parameters at this time are the results of the original model for flight reliable perception of the physical conditions of the drone swing amplitude. In order to improve the stability of the flight reliable perception results, the embodied intelligent data processing platform can construct the physical conditions of the drone swing amplitude as the observed drone swing amplitude physical condition edge computing gateway, and randomly combine the drone swing amplitude physical conditions in the observed drone swing amplitude physical condition edge computing gateway, so that the fusion of information content can be achieved in the combination. The present invention processes the physical conditions of the drone swing amplitude for flight reliable perception, ensures the flight safety of the drone, and saves hardware costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0044] Figure 1 The present invention is a flowchart of a method for embodied intelligent and reliable perception of a UAV based on collaborative evolution in one embodiment. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0046] In one embodiment, Figure 1 As shown, a UAV embodied intelligent reliable perception method based on co-evolution is provided, comprising the following steps:

[0047] Step A102: Integrate the observed UAV crash risk factors, process the observed UAV crash risk factors using a cluster analysis algorithm, and obtain the physical conditions of the UAV swing amplitude.

[0048] Preferably, a cluster analysis algorithm is used to classify the risk factors of drone crashes according to different set conditions, such as flight speed, flight distance, flight altitude, etc., and the risk factors of drone crashes under different environmental conditions. Different categories represent different crash probabilities, and the crash probability here is based on historical flight record statistics.

[0049] Step A104: Use the set co-evolution model to perform flight reliable perception processing on the physical condition of the UAV's swing amplitude, obtain the flight attitude reliable perception parameters corresponding to the physical condition of the UAV's swing amplitude, and use the UAV's swing amplitude physical condition and the flight attitude reliable perception parameters to build an edge computing gateway for the observed UAV's swing amplitude physical condition.

[0050] Preferably, the flight attitude reliable perception parameter can be data obtained by processing the physical condition of the UAV swing amplitude by the co-evolution model, and can match the information of the physical condition of the UAV swing amplitude in a specific flight environment.

[0051] The set co-evolution model is used to perform flight reliable perception processing on the physical conditions of the UAV swing amplitude, and the expression is:

[0052]

[0053] in, represents the reliable perception parameter of flight attitude, G represents the weight of the crash affected by human factors, U represents the weight of the crash affected by non-human factors, ξ represents the probability of occurrence of each human factor affecting the crash, Te represents the distribution matrix of the types of human factors affecting the crash, ω represents the probability of occurrence of each non-human factor affecting the crash, Ti represents the distribution matrix of the types of non-human factors affecting the crash, K represents the reliable perception coefficient, represents the embodied intelligence observation error for each physical condition, X j , Y j , Z j Respectively represent the flight attitude roll angle, pitch angle, and yaw angle;

[0054] Step A106: The edge computing gateway of the physical condition of the swing amplitude of the observed drone generates a boundary threshold of the physical condition of the swing amplitude of the observed drone, and sends the boundary threshold of the physical condition of the swing amplitude of the observed drone to the embodied intelligent data processing platform.

[0055] Preferably, the boundary threshold of the physical condition of the swing amplitude of the observed drone can be a boundary threshold obtained by combining several physical conditions of the swing amplitude of the drone. The boundary threshold of the physical condition of the swing amplitude of the observed drone can be composed of physical conditions of the swing amplitude of the drone derived from different risk factors of the observed drone crash. The embodied intelligence data processing platform is a data processing platform combined with artificial intelligence. Embodied intelligence refers to the ability of an intelligent system or machine to interact with the environment in real time through perception and interaction. It can be simply understood as robots of various forms, allowing them to perform various tasks in a real physical environment to complete the evolution of artificial intelligence. The robot in the present invention specifically refers to a drone.

[0056] Step A108: Send information according to the instruction of the embodied intelligent data processing platform to integrate the reliable perception information of the flight altitude of the physical condition boundary threshold of the swing amplitude of the observed UAV.

[0057] The command sending information may refer to the confirmation information input by the operator during the operation for the physical condition boundary threshold of the swing amplitude of the observed UAV.

[0058] Step A110, respectively process the flight attitude reliable perception parameters and the flight altitude reliable perception information using a convolutional neural network, adjust the parameters of the observed UAV swing amplitude physical condition edge computing gateway according to the flight attitude reliable perception parameters mapped to the UAV swing amplitude physical condition at the same altitude and the flight influence weights of the flight altitude reliable perception information, and iterate the co-evolution model using the observed UAV swing amplitude physical condition edge computing gateway after parameter adjustment.

[0059] The co-evolution model is a specific flight reliability perception optimization model that is used to identify potential human or non-human influence information in complex flight environments, and use the results of the identification to replace and compensate for error interference, thereby ensuring the stability of the analysis attitude. The advantage of the co-evolution model is that it can be used to quickly process multiple risk factors with high efficiency, without the need for large amounts of data calculation, and saving hardware costs.

[0060] The convolutional neural network uses convolution kernels to perform feature extraction on the reliable perception parameters of the flight attitude and the reliable perception information of the flight altitude, extracts noise-free reliable perception parameters, and uses a fully connected layer to integrate the two-dimensional feature vectors of the noise-free reliable perception parameters into one-dimensional vectors. Each one-dimensional vector is mapped to a flight impact weight.

[0061] The co-evolution model in the present invention obtains the physical conditions of the drone swing amplitude contained in the observed drone crash risk factors by segmenting and processing the features of the observed drone crash risk factors. At this time, the physical conditions of the drone swing amplitude may include a single human or non-human factor. Then, the initial co-evolution model can be applied to the physical conditions of the drone swing amplitude for flight reliable perception, and obtain flight attitude reliable perception parameters. At this time, the flight attitude reliable perception parameters are the results of the original model for flight reliable perception of the physical conditions of the drone swing amplitude. In order to improve the stability of the flight reliable perception results, the embodied intelligent data processing platform can construct the physical conditions of the drone swing amplitude as the observed drone swing amplitude physical condition edge computing gateway, and randomly combine the drone swing amplitude physical conditions in the observed drone swing amplitude physical condition edge computing gateway, so that the fusion of information content can be achieved in the combination. The present invention processes the physical conditions of the drone swing amplitude for flight reliable perception, thereby ensuring the flight safety of the drone.

[0062] In the above-mentioned UAV embodied intelligent reliable perception method based on collaborative evolution, reasonable deduction is performed in combination with the technical features in the embodiment, which can achieve the following beneficial effects of solving the technical problems raised in the background technology:

[0063] By segmenting and processing the features of the observed UAV crash risk factors, the physical conditions of the UAV swing amplitude included in the observed UAV crash risk factors are obtained. At this time, the physical conditions of the UAV swing amplitude can include a single human or non-human factor. Then, the initial co-evolution model can be applied to the physical conditions of the UAV swing amplitude for flight reliability perception, and the flight attitude reliable perception parameters are obtained. At this time, the flight attitude reliable perception parameters are the results of the original model for flight reliability perception of the physical conditions of the UAV swing amplitude. In order to improve the stability of the flight reliability perception results, the embodied intelligent data processing platform can construct the physical conditions of the UAV swing amplitude as the edge computing gateway of the observed UAV swing amplitude physical conditions, and randomly combine the UAV swing amplitude physical conditions in the edge computing gateway of the observed UAV swing amplitude physical conditions. Subsequently, the boundary threshold of the observed UAV swing amplitude physical conditions can be sent to the embodied intelligent data processing platform, and the meaning of these observed UAV swing amplitude physical conditions can be reliably perceived, so as to analyze and process the interfering strange and special UAV swing amplitude physical conditions. Finally, the flight height reliable perception information can be combined with the flight attitude reliable perception parameters, so as to adjust the parameters of the co-evolution model, improve the reliability of the co-evolution model data, and finally improve the stability of the model.

[0064] Preferably, the sending of parameters for reliable perception of flight control is achieved in a 5G manner.

[0065] Integrate a plurality of said flight altitude reliable perception information mapped with the physical condition of the swing amplitude of the observed UAV;

[0066] The flight influence weights in different climate environments of the flight attitude reliable perception parameters and the flight altitude reliable perception information respectively integrated with the physical conditions of the swing amplitude of the observed UAV;

[0067] If the mean value of the flight impact weight in different climate environments exceeds the set weight safety parameter, it is determined that the flight attitude reliable perception parameter is interfered by environmental factors, and the flight altitude reliable perception information in different climate environments is normalized to obtain standardized flight reliable perception information, and the physical condition edge computing gateway of the swing amplitude of the observed UAV is adjusted with the standardized flight reliable perception information parameters.

[0068] In implementation, the role of the edge computing gateway is similar to that of the cloud, which can be used for data storage, analysis, and running of model codes. The edge computing gateway is connected to the drone in real time via 5G signals.

[0069] A coefficient matrix corresponding to the co-evolution model after integrated parameter adjustment and the physical condition of the UAV swing amplitude;

[0070] The error sources of the coefficient matrix and the flight attitude reliable perception parameters are identified to identify interference substitution parameter pairs in the physical conditions of the UAV swing amplitude, wherein the interference substitution parameter pairs include error interference parameters for identifying the flight attitude reliable perception parameters and allowable interference parameters for identifying the coefficient matrix, and the error interference parameters correspond to the allowable interference parameters.

[0071] The interference substitution parameter is used to construct interference substitution data, and the set Transformer model is trained using the interference substitution data to obtain an interference substitution model, which is used to perform interference substitution on the physical condition of the swing amplitude of the drone.

[0072] The interference substitution model is expressed as:

[0073]

[0074] Among them, D s represents interference replacement data, P1…P n represents the interference substitution parameter matrix, R 11 …R nn represents the influence weight of each interference substitution parameter, F1…F nIndicates whether the interference substitution parameter can be replaced by a matrix. The judgment basis is whether the influence weight of each interference substitution parameter exceeds the preset value. When it is greater than the preset value, it is 1, and when it is less than the preset value, it is 0. j represents the Transformer encoding-decoding efficiency, H c Represents the input sequence of Transformer, i represents the model training count, and I represents the preset total number of model training times.

[0075] The interference substitution model is applied to preprocess the physical condition of the UAV swing amplitude so that the interference in the physical condition of the UAV swing amplitude can be compensated.

[0076] The observed UAV swing amplitude physical condition edge computing gateway is updated using the set For loop + Update statement. The update process is used to delete the observed UAV swing amplitude physical condition in the observed UAV swing amplitude physical condition edge computing gateway. The observed UAV swing amplitude physical condition includes the UAV swing amplitude physical condition whose error interference parameter is less than the set parameter range.

[0077] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the indications of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same altitude, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0078] The embodiment of the present application also provides a drone embodied intelligent and reliable perception system based on co-evolution. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of a drone embodied intelligent and reliable perception system based on co-evolution provided below can refer to the limitations of a drone embodied intelligent and reliable perception method based on co-evolution above, and will not be repeated here.

[0079] In one embodiment, an operating system of a UAV embodied intelligent reliable perception method based on co-evolution is provided, comprising: a cluster analysis unit, an embodied intelligent flight reliable perception unit, a physical condition sending unit for the swing amplitude of an observed UAV, a boundary threshold information flight reliable perception unit, and a flight reliable perception model iteration unit, wherein:

[0080] A cluster analysis unit, used to integrate the observed UAV crash risk factors, process the observed UAV crash risk factors using a cluster analysis algorithm, and obtain the UAV swing amplitude physical conditions;

[0081] The embodied intelligent flight reliable perception unit is used to perform flight reliable perception processing on the physical condition of the UAV swing amplitude by using the set co-evolution model, obtain flight attitude reliable perception parameters corresponding to the physical condition of the UAV swing amplitude, and construct an edge computing gateway for the physical condition of the observed UAV swing amplitude by using the physical condition of the UAV swing amplitude and the reliable perception parameters of the flight attitude;

[0082] The observed UAV swing amplitude physical condition sending unit is used to generate the observed UAV swing amplitude physical condition boundary threshold value by the observed UAV swing amplitude physical condition edge computing gateway, and send the observed UAV swing amplitude physical condition boundary threshold value to the embodied intelligent data processing platform;

[0083] A flight reliability perception unit for boundary threshold information, used to send information according to the instruction of the embodied intelligent data processing platform, and integrate the flight altitude reliable perception information of the boundary threshold of the physical condition of the swing amplitude of the observed UAV;

[0084] The flight reliable perception model iteration unit is used to process the flight attitude reliable perception parameters and the flight altitude reliable perception information respectively by using a convolutional neural network, adjust the parameters of the observed UAV swing amplitude physical condition edge computing gateway according to the flight attitude reliable perception parameters mapped to the UAV swing amplitude physical condition at the same altitude and the flight influence weight of the flight altitude reliable perception information, and iterate the co-evolution model using the observed UAV swing amplitude physical condition edge computing gateway after parameter adjustment.

[0085] Furthermore, the flight reliability perception model iteration unit includes:

[0086] A flight altitude reliable perception information unit, used to integrate a plurality of said flight altitude reliable perception information mapped with the physical conditions of the swing amplitude of the observed UAV;

[0087] A flight influence weight unit, used to respectively integrate the flight influence weights in different climate environments of the flight attitude reliable perception parameters and the flight altitude reliable perception information mapped with the physical conditions of the swing amplitude of the observed UAV;

[0088] The information normalization unit is used to determine that the flight attitude reliable perception parameter is interfered by environmental factors if the mean value of the flight impact weight in different climate environments exceeds the set weight safety parameter, and normalize the flight altitude reliable perception information in different climate environments to obtain standardized flight reliable perception information, and adjust the physical condition edge computing gateway of the swing amplitude of the observed UAV with the standardized flight reliable perception information parameters.

[0089] Furthermore, after the flight reliability perception model iteration unit, it also includes:

[0090] A coefficient matrix unit, used for integrating the coefficient matrix corresponding to the co-evolution model after parameter adjustment and the physical condition of the swing amplitude of the UAV;

[0091] An interference substitution parameter pair unit is used to identify the error source of the coefficient matrix and the flight attitude reliable perception parameter, and identify the interference substitution parameter pair in the physical condition of the UAV swing amplitude, wherein the interference substitution parameter pair includes an error interference parameter for identifying the flight attitude reliable perception parameter, and also includes an allowable interference parameter for identifying the coefficient matrix, wherein the error interference parameter corresponds to the allowable interference parameter;

[0092] The interference substitution model unit is used to construct interference substitution data using the interference substitution parameter pair, and train the set Transformer model using the interference substitution data to obtain an interference substitution model, wherein the interference substitution model is used to perform interference substitution on the physical condition of the UAV swing amplitude.

[0093] Furthermore, before the embodied intelligent flight reliable perception unit, it also includes:

[0094] A data preprocessing unit is used to apply the interference substitution model to preprocess the physical condition of the UAV swing amplitude so that the interference in the physical condition of the UAV swing amplitude can be compensated.

[0095] Furthermore, before the physical condition sending unit of the swing amplitude of the observed drone, it also includes:

[0096] A data update unit is used to update the observed UAV swing amplitude physical condition edge computing gateway using a set For loop + Update statement, wherein the update process is used to delete the observed UAV swing amplitude physical condition in the observed UAV swing amplitude physical condition edge computing gateway, and the observed UAV swing amplitude physical condition includes the UAV swing amplitude physical condition whose error interference parameter is less than the set parameter range.

[0097] Each unit in the above-mentioned cooperative evolution-based embodied intelligent reliable perception system for unmanned aerial vehicles can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned units can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above-mentioned units.

[0098] Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A UAV embodied intelligent and reliable perception method based on co-evolution, characterized in that: The following steps are involved: Step 1: Integrate the observed UAV crash risk factors, process the observed UAV crash risk factors through cluster analysis algorithm, and obtain the physical conditions of the UAV swing amplitude; Step 2: Use the set co-evolution model to perform flight reliable perception processing on the physical condition of the UAV swing amplitude to obtain the flight attitude reliable perception parameter corresponding to the physical condition of the UAV swing amplitude, and the expression is: in, represents the reliable perception parameter of flight attitude, G represents the weight of the crash affected by human factors, U represents the weight of the crash affected by non-human factors, ξ represents the probability of occurrence of each human factor affecting the crash, Te represents the distribution matrix of the types of human factors affecting the crash, ω represents the probability of occurrence of each non-human factor affecting the crash, Ti represents the distribution matrix of the types of non-human factors affecting the crash, K represents the reliable perception coefficient, represents the embodied intelligence observation error for each physical condition, X j , Y j , Z j Respectively represent the flight attitude roll angle, pitch angle, and yaw angle; Step 3, using the physical condition of the swing amplitude of the drone and the reliable perception parameters of the flight attitude to build an edge computing gateway for the physical condition of the swing amplitude of the observed drone; the reliable perception parameters of the flight attitude are the embodied intelligent flight reliable perception information integrated with the reliable flight perception of the physical condition of the swing amplitude of the drone through the co-evolution model; Step 4: The observed UAV swing amplitude physical condition edge computing gateway generates the observed UAV swing amplitude physical condition boundary threshold, and sends the observed UAV swing amplitude physical condition boundary threshold to the embodied intelligent data processing platform; Step 5, according to the command information sent by the embodied intelligent data processing platform, integrate the flight altitude reliable perception information of the boundary threshold of the physical condition of the swing amplitude of the observed UAV, wherein the flight altitude reliable perception information is the flight control reliable perception information input by the operator of the embodied intelligent data processing platform during the command sending process; Step 6, respectively processing the flight attitude reliable perception parameters and the flight altitude reliable perception information through a convolutional neural network, adjusting the parameters of the observed UAV swing amplitude physical condition edge computing gateway according to the flight attitude reliable perception parameters mapped to the UAV swing amplitude physical condition at the same altitude and the flight influence weight of the flight altitude reliable perception information, and iterating the co-evolution model using the observed UAV swing amplitude physical condition edge computing gateway after parameter adjustment; Step 7, integrating the coefficient matrix corresponding to the co-evolution model after parameter adjustment and the physical condition of the swing amplitude of the UAV; Identify the error source of the coefficient matrix and the flight attitude reliable perception parameter, and identify the interference substitution parameter pair in the physical condition of the UAV swing amplitude, wherein the interference substitution parameter pair includes an error interference parameter for identifying the flight attitude reliable perception parameter, and also includes an allowable interference parameter for identifying the coefficient matrix, wherein the error interference parameter corresponds to the allowable interference parameter; Using the interference substitution parameter pair to construct interference substitution data, using the interference substitution data to train a set Transformer model to obtain an interference substitution model, wherein the interference substitution model is used to perform interference substitution on the physical condition of the swing amplitude of the drone; The interference substitution model is expressed as: Among them, D s represents interference replacement data, P1…P n represents the interference substitution parameter matrix, R 11 …R nn represents the influence weight of each interference substitution parameter, F1…F n Indicates whether the interference substitution parameter can be replaced by a matrix. The judgment basis is whether the influence weight of each interference substitution parameter exceeds the preset value. When it is greater than the preset value, it is 1, and when it is less than the preset value, it is 0. j represents the Transformer encoding-decoding efficiency, H c Represents the input sequence of Transformer, i represents the model training count, and I represents the preset total number of model training times.

2. The method for reliable perception of drone-embodied intelligence based on collaborative evolution according to claim 1 is characterized in that: In step 6, The convolutional neural network uses convolution kernels to perform feature extraction on the reliable perception parameters of flight attitude and the reliable perception information of flight altitude, extracts noise-free reliable perception parameters, and uses a fully connected layer to integrate the two-dimensional feature vectors of the noise-free reliable perception parameters into one-dimensional vectors, and each one-dimensional vector is mapped to a flight impact weight.

3. The method for reliable perception of drone-embodied intelligence based on collaborative evolution according to claim 1 is characterized in that: In the step 6, iterating the co-evolution model using the edge computing gateway of the observed UAV swing amplitude physical condition after parameter adjustment includes: integrating a plurality of the flight altitude reliable perception information mapped with the observed UAV swing amplitude physical condition; The flight influence weights in different climate environments of the flight attitude reliable perception parameters and the flight altitude reliable perception information respectively integrated with the physical conditions of the swing amplitude of the observed UAV; If the mean value of the flight impact weight in different climate environments exceeds the set weight safety parameter, it is determined that the flight attitude reliable perception parameter is interfered by environmental factors, and the flight altitude reliable perception information in different climate environments is normalized to obtain standardized flight reliable perception information, and the physical condition edge computing gateway of the swing amplitude of the observed UAV is adjusted with the standardized flight reliable perception information parameters.

4. The method for reliable perception of drone-embodied intelligence based on collaborative evolution according to claim 1 is characterized in that: Between step 1 and step 2 also include: The interference substitution model is applied to preprocess the physical condition of the UAV swing amplitude so that the interference in the physical condition of the UAV swing amplitude can be compensated.

5. The method for reliable perception of drone-embodied intelligence based on collaborative evolution according to claim 1 is characterized in that: Also include before step 4: The observed UAV swing amplitude physical condition edge computing gateway is updated using the set For loop + Update statement. The update process is used to delete the observed UAV swing amplitude physical condition in the observed UAV swing amplitude physical condition edge computing gateway. The observed UAV swing amplitude physical condition includes the UAV swing amplitude physical condition whose error interference parameter is less than the set parameter range.

6. A UAV embodied intelligent and reliable perception method based on collaborative evolution according to any one of claims 1 to 5, characterized in that: This is achieved through a UAV-embodied intelligent and reliable perception system based on co-evolution, which includes: A cluster analysis unit, used to integrate the observed UAV crash risk factors, process the observed UAV crash risk factors using a cluster analysis algorithm, and obtain the UAV swing amplitude physical conditions; The embodied intelligent flight reliable perception unit is used to perform flight reliable perception processing on the physical condition of the UAV swing amplitude by using the set co-evolution model, obtain flight attitude reliable perception parameters corresponding to the physical condition of the UAV swing amplitude, and construct an edge computing gateway for the physical condition of the observed UAV swing amplitude by using the physical condition of the UAV swing amplitude and the flight attitude reliable perception parameters, wherein the flight attitude reliable perception parameters are embodied intelligent flight reliable perception information integrated with flight reliable perception of the physical condition of the UAV swing amplitude by using the co-evolution model; The observed UAV swing amplitude physical condition sending unit is used to generate the observed UAV swing amplitude physical condition boundary threshold value by the observed UAV swing amplitude physical condition edge computing gateway, and send the observed UAV swing amplitude physical condition boundary threshold value to the embodied intelligent data processing platform; A flight reliability perception unit for boundary threshold information is used to send information according to the instruction of the embodied intelligent data processing platform, and integrate the flight altitude reliability perception information of the boundary threshold of the physical condition of the swing amplitude of the observed UAV, and the flight altitude reliability perception information is the flight control reliability perception information input by the operator of the embodied intelligent data processing platform during the instruction sending process; The flight reliable perception model iteration unit is used to process the flight attitude reliable perception parameters and the flight altitude reliable perception information respectively by using a convolutional neural network, adjust the parameters of the observed UAV swing amplitude physical condition edge computing gateway according to the flight attitude reliable perception parameters mapped to the UAV swing amplitude physical condition at the same altitude and the flight influence weight of the flight altitude reliable perception information, and iterate the co-evolution model using the observed UAV swing amplitude physical condition edge computing gateway after parameter adjustment.

7. The method for reliable perception of drone-embodied intelligence based on collaborative evolution according to claim 6 is characterized in that: The flight reliability perception model iteration unit includes: A flight altitude reliable perception information unit, used to integrate a plurality of said flight altitude reliable perception information mapped with the physical conditions of the swing amplitude of the observed UAV; A flight influence weight unit, used to respectively integrate the flight influence weights in different climate environments of the flight attitude reliable perception parameters and the flight altitude reliable perception information mapped with the physical conditions of the swing amplitude of the observed UAV; The information normalization unit is used to determine that the flight attitude reliable perception parameter is interfered by environmental factors if the mean value of the flight impact weight in different climate environments exceeds the set weight safety parameter, and normalize the flight altitude reliable perception information in different climate environments to obtain standardized flight reliable perception information, and adjust the physical condition edge computing gateway of the swing amplitude of the observed UAV with the standardized flight reliable perception information parameters.

Images