Intelligent laying method and system for aircraft system pipeline

Abstract

The invention discloses an intelligent laying method and system for an aircraft system pipeline. The method comprises the following steps: acquiring three-dimensional data and positioning information of the aircraft system pipeline; and generating an optimal pipeline laying path. And the pipeline position is detected in real time, the optimal pipeline laying path is superposed on the pipeline position detected in real time, and the optimal installation position of the pipeline clamp is projected in real time. According to the invention, based on advanced technologies such as depth vision, intelligent path planning and augmented reality guidance, intelligentization and automation of the whole process of aircraft system pipeline laying are realized. The method has the remarkable advantages in the aspects of improving efficiency, enhancing precision, reducing cost, improving user experience and the like, a comprehensive solution is provided for intelligent upgrading of the aeronautical manufacturing industry, and the method has wide application prospects and important practical value.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of aircraft system pipeline laying, and particularly relates to an aircraft system pipeline intelligent laying method and system. BACKGROUND

[0002] With the rapid development of the aviation manufacturing industry, modern aircraft structures are increasingly complex. As the "blood vessels" and "nerves" of the aircraft, aircraft system pipeline systems play a crucial role in ensuring the normal operation of the aircraft. Traditional aircraft system pipeline laying methods mainly rely on manual operation, which not only consumes time and effort, but also is prone to assembly errors, making it difficult to meet the high requirements of modern aviation manufacturing on assembly precision and quality.

[0003] In recent years, some semi-automatic pipeline laying systems have been gradually applied in the aviation manufacturing industry. These systems usually use mechanical arms in combination with simple vision systems to position and install pipelines. However, this method has poor adaptability in complex environments and is difficult to handle the narrow and winding spaces inside the aircraft cabin. At the same time, due to the lack of intelligent path planning capabilities, these systems often require manual path presetting and cannot dynamically adjust according to actual conditions, resulting in suboptimal pipeline layout and potential interference risks. On the other hand, although three-dimensional modeling technology is widely used in the design stage of aircrafts, there are still significant challenges in converting virtual design into actual installation. Existing augmented reality technology can assist the assembly process to some extent, but due to the lack of real-time environmental perception and intelligent decision-making capabilities, its guidance effect is limited and it is difficult to adapt to complex and variable actual assembly environments. In addition, existing pipeline laying systems generally have low system integration and poor coordination between modules. For example, the path planning module and the actual installation guidance module are often fragmented, making it difficult to achieve real-time interaction and feedback of information, which leads to a disconnect between theoretical design and actual installation, affecting the overall assembly quality.

[0004] For example, the existing Chinese patent CN117763770A discloses an aircraft engine pipeline laying method, which includes the following steps: constructing an aircraft engine pipeline laying space mathematical model, generating a laying map according to the mathematical model; searching for an initial path using the A* algorithm according to the laying map, taking the starting point as the initial node, and using an evaluation function to evaluate the candidate points around the node, the candidate point with the smallest evaluation function value being taken as the next node; using a PSO optimization algorithm to optimize the initial path, so that the curved part of the path corresponding to the pipeline has processability.

[0005] For example, the existing Chinese patent CN117113603A discloses an aero-engine external pipeline simulation laying method and system, the method comprising the following steps: generating an engine casing model and a model of obstacles on the engine casing; determining the starting point and the ending point position of the engine casing external pipeline; generating a pipeline laying map; calculating the pipeline path by using a pipeline path planning algorithm, and performing post-processing on the generated pipeline path, the post-processing comprising: removing redundant points, removing redundant turns, stretching the pipeline to make it processable, and installing edge adsorption; generating spline curve nodes according to the path nodes obtained after the post-processing; and generating a center line and a pipeline according to the spline curve nodes. SUMMARY

[0006] The purpose of the present application is to provide an aircraft system pipeline intelligent laying method and system, aiming to solve the problems of insufficient assembly precision, low efficiency and high interference risk in the existing aircraft system pipeline laying.

[0007] The present application is mainly realized by the following technical solutions:

[0008] An aircraft system pipeline intelligent laying method, comprising the following steps:

[0009] Step S1: acquiring three-dimensional data and positioning information of the aircraft system pipeline;

[0010] Step S11: collecting three-dimensional data of the aircraft system pipeline; at the same time, acquiring spatial pose information of the on-board inertial sensor;

[0011] Step S12: calculating the three-dimensional data of the aircraft system pipeline in the on-board inertial sensor coordinate system based on the spatial pose information;

[0012] Step S13: converting the three-dimensional data in the on-board inertial sensor coordinate system into point cloud position information in the world coordinate system;

[0013] Step S2: generating an optimal pipeline laying path;

[0014] Step S21: based on the point cloud position information, generating a continuous radiation field by using a radiation field generation algorithm, constructing a three-dimensional radiation field model; dividing the three-dimensional space into multiple grids and calculating the probability density of each grid; based on the probability density, combining the gray value G(x, i) of the point, calculating the grid feature vector F(i) to input the graph neural network;

[0015] Step S22: based on the graph neural network, extracting the grid features and calculating the probability of the pipeline ending point;

[0016] Step S23: generating a pipeline laying path by using a deep deterministic policy gradient algorithm for path planning;

[0017] Step S24: obtaining the optimal pipeline laying path based on the pipeline interference constraint condition;

[0018] Step S3: detecting the pipeline position in real time, superimposing the optimal pipeline laying path on the pipeline position detected in real time, and projecting the optimal installation position of the pipeline clamp in real time.

[0019] In order to better realize the present application, further, in step S13, the Kalman filtering algorithm is used for coordinate conversion, wherein the state equation and the observation equation are as follows:

[0020] X k =F k X k-1 +B k U k +w k ;

[0021] Z k =H k X k +v k ;

[0022] Wherein: X k is a state vector;

[0023] X k-1 is a state vector of the system at k-1 time, containing the position, speed and attitude information measured by the inertial measurement unit;

[0024] F k is a state transition matrix;

[0025] B k is a control input matrix;

[0026] U k is a control vector;

[0027] w k is a process noise;

[0028] Z k is an observation vector;

[0029] H k is an observation matrix;

[0030] v k is an observation noise.

[0031] In order to better realize the present application, further, in step S21, the radiation field generation algorithm is as follows:

[0032]

[0033] Wherein, F(x, y, z) represents the radiation field intensity at the space point (x, y, z).

[0034] (x i , y i , z i ) is the coordinate of the i-th sampling point;

[0035] N is the number of sampling points;

[0036] w i is the weight of the i-th sampling point;

[0037] φ() is a radial basis function;

[0038] Calculate the probability density of each grid:

[0039]

[0040] where: x is a point in the point cloud,

[0041] c i represents the center of grid i,

[0042] c i represents the center of grid j,

[0043] σ is an adjustable parameter;

[0044] The gray value G(x, i) of the point is:

[0045] G(x, i) = I(x) * P(x | i);

[0046] The feature vector F(i) of each grid is:

[0047]

[0048] In order to better realize the present application, further, the step S22 comprises the following steps:

[0049] Step A1: Calculate the relationship features of each node and the surrounding nodes, and generate a node feature vector;

[0050]

[0051] where: R(p, q) is the relationship feature of each node and the surrounding nodes;

[0052] ||p-q||2 represents the Euclidean distance between nodes p and q;

[0053] and respectively represent the i-th adjacent node of node p and node q;

[0054] K is the number of adjacent nodes considered;

[0055] G(p, q) is a Gaussian kernel function:

[0056]

[0057] Step A2: Calculate the weight relationship between nodes, and generate edge feature vectors;

[0058]

[0059] Where: W(p, q) is the weight relationship between node p and node q;

[0060] Beta is an adjustable parameter;

[0061] N(p) represents the neighborhood set of node p;

[0062] ||p-v||2 is the Euclidean distance between node p and node v;

[0063] Step A3: Aggregate node features and edge features, and update node representation;

[0064]

[0065] Where: W (k) And b( k) Are learnable parameters;

[0066] is the feature representation of node u at the kth layer;

[0067] is the feature representation of node v at the k+1th layer;

[0068] Alpha vu Is the attention coefficient;

[0069] Sigma() is an activation function;

[0070] N(v) is the neighborhood set of node v;

[0071] Step A4: Calculate the probability of each node becoming a pipeline endpoint.

[0072] In order to better realize the present application, further, in the step S22, the formula of the graph neural network is:

[0073]

[0074] Where: is the feature representation of node v at the kth layer;

[0075] Sigma() is an activation function;

[0076] W (k) And b( k) Are learnable parameters;

[0077] is the feature representation of node u at the k-1 layer;

[0078] N(v) is the neighborhood set of node v;

[0079] AGGREGATE (k) is the aggregation function:

[0080]

[0081] where: a() is the attention score function;

[0082] α vu is the attention coefficient;

[0083] u is the node within the neighborhood, representing the grid position adjacent to the current node. The features of each neighbor node are weighted aggregated according to the attention weight, forming the new feature representation of the node. This mechanism allows the network to adaptively aggregate adjacent information, more accurately capturing the spatial structure;

[0084] is the feature representation of node v at the k-1 layer;

[0085] is the feature vector (or hidden state) of neighbor node w in the (k-l) layer of the graph neural network.

[0086] In order to better achieve the present application, further, step S24 comprises the following steps:

[0087] Step B1: Discretize the pipe surface into a series of point sets {p i}, then calculate the minimum distance from each point to the obstacle:

[0088]

[0089] where o represents the obstacle point set;

[0090] Step B2: Construct a continuous distance field function:

[0091]

[0092] where w i (x) is the weight function:

[0093]

[0094] Step B2: If D(x) < d threshold , it is considered that there is a risk of collision;

[0095] where: dthreshold is a preset distance threshold.

[0096] To better achieve the present application, further, the step S3 comprises the following steps:

[0097] Step S31: Real-time detection of pipeline position by laser sensor, and generation of pipeline position data;

[0098] Step S32: Superimposing the optimal pipeline laying path on the pipeline position data, and dynamically adjusting the rendering parameters by using an adaptive rendering algorithm to generate an augmented reality scene;

[0099] The adaptive rendering algorithm is:

[0100] I(x,y)=k a I a +k d (L·N)I d +k s (R·V) n I s ;

[0101] wherein I(x,y) represents the final rendering intensity of pixel (x,y);

[0102] k a , k d and k s are the ambient light, diffuse reflection and specular reflection coefficients respectively;

[0103] I a , I d and I s are the intensities of ambient light, diffuse reflection light and specular reflection light respectively;

[0104] L is a light source direction vector;

[0105] N is a surface normal vector;

[0106] R is a reflected light direction vector;

[0107] V is a line of sight direction vector;

[0108] n is a specular reflection index;

[0109] Step S33: Calculation of the optimal installation position of the clamp in the optimal pipeline laying path, wherein the number of clamps is minimized under the premise of meeting the mechanical constraints;

[0110] Define a state transition function f(i) representing the minimum number of clamps required from the starting point to the i-th point:

[0111]

[0112] Wherein, D max is the maximum allowable distance between the clamps;

[0113] D(i,j) is the distance between any point i(x i , y i , z i ) and point j(x j , y j , z j ) on the path;

[0114] Step S34: using multiple contour line simulation techniques, projecting the optimal installation position of the pipeline clamp in real time through the AR glasses.

[0115] The present application is mainly realized by the following technical solutions:

[0116] An aircraft system pipeline intelligent laying system for realizing the above-mentioned aircraft system pipeline intelligent laying method, comprising a depth vision subsystem, an intelligent path planning subsystem and an augmented reality guidance subsystem connected in sequence from front to back;

[0117] The depth vision subsystem is used to acquire three-dimensional data and positioning information of the aircraft system pipeline, generate a pipeline three-dimensional model, and determine point cloud position information of point cloud data in a world coordinate system according to an inertial measurement unit coordinate system;

[0118] The intelligent path planning subsystem is used to generate a three-dimensional radiation field model based on the three-dimensional model and the point cloud position information, and plan an optimal pipeline laying path in the three-dimensional radiation field model by using a graph neural network and a reinforcement learning algorithm;

[0119] The augmented reality guidance subsystem is used to detect the pipeline position in real time, superimpose the optimal pipeline laying path on the real-time detected pipeline position, and project the optimal installation position of the pipeline clamp in real time through the augmented reality glasses.

[0120] In order to better realize the present application, further, the depth vision subsystem comprises a laser radar module, an inertial measurement unit module and a coordinate conversion module connected in sequence from front to back, and a photoelectric encoder, which is respectively in communication connection with the laser radar module and the inertial measurement unit module;

[0121] The laser radar module is used to acquire three-dimensional data and positioning information of the aircraft system pipeline;

[0122] The inertial measurement unit module is used to acquire space pose information of an airborne inertial sensor and calculate the position of the three-dimensional data of the aircraft system pipeline in the airborne inertial sensor coordinate system;

[0123] The photoelectric encoder is used to acquire the scanning frequency of the laser radar module and record the current timestamp of the inertial measurement unit module.

[0124] A coordinate conversion module is configured to convert three-dimensional data in an on-board inertial sensor coordinate system into point cloud position information in a world coordinate system.

[0125] To better achieve the present application, further, the intelligent path planning subsystem comprises, in sequence from front to back, a radiation field generation module, a graph neural network module, a reinforcement learning module and a pipeline interference avoidance module.

[0126] The radiation field generation module is configured to divide a three-dimensional space into a plurality of grids based on a three-dimensional model and point cloud position information, calculate a probability density of each grid, and generate a three-dimensional radiation field model.

[0127] The graph neural network module is configured to receive the three-dimensional radiation field model, extract a feature vector of each grid, and calculate a probability of each grid as a pipeline end point.

[0128] The reinforcement learning module is configured to receive the probability information of the grid and select an action strategy based on an epsilon-greedy algorithm, and generate a pipeline laying path through iterative optimization.

[0129] The pipeline interference avoidance module is configured to receive the pipeline laying path, and obtain an optimal pipeline laying path based on interference constraints.

[0130] To better achieve the present application, further, the augmented reality guidance subsystem comprises, in sequence from front to back, a laser sensor module, a simulation rendering module, a clamp position calculation module and an AR projection module.

[0131] The laser sensor module is configured to detect a pipeline position in real time and generate pipeline position data.

[0132] The simulation rendering module is configured to receive the optimal pipeline laying path, superimpose the optimal pipeline laying path on the pipeline position data, and generate an augmented reality scene.

[0133] The clamp position calculation module is configured to determine a clamp position in the optimal pipeline laying path according to a preset distance threshold, and generate clamp position data.

[0134] The AR projection module is configured to receive the clamp position data and project a best installation position of a pipeline clamp in real time through AR glasses.

[0135] The present application has the following advantages:

[0136] The present application is based on advanced technologies such as deep vision, intelligent path planning and augmented reality guidance, and can realize the intelligentization and automation of the whole process of aircraft system pipeline laying. The present application has significant advantages in improving efficiency, enhancing precision, reducing cost, and improving user experience, and provides a comprehensive solution for the intelligent upgrading of the aviation industry, which has broad application prospects and important practical value.

[0137] From the overall architecture, the present application realizes the seamless integration of three subsystems of deep vision, intelligent path planning and augmented reality guidance, forming a closed-loop intelligent laying process. Specifically, the various modules of the present application realize high coordination and complementation. For example, the deep vision subsystem provides accurate spatial information for path planning, and the results of path planning are fed back to the operator in real time through the AR system. This close information interaction ensures the high consistency of virtual design and actual installation, effectively solving the "design-installation" disconnection problem in traditional methods. Specifically, the present application introduces an augmented reality guidance subsystem. By combining real-time environment perception and high-precision AR projection technology, the system provides intuitive and accurate installation guidance for the operator. In particular, the use of multiple contour simulation technology not only provides accurate spatial positioning information, but also effectively solves the display problem of traditional AR systems in complex lighting environments. This greatly reduces human operation errors and improves installation accuracy and efficiency. The highly integrated system design of the present application greatly improves the overall efficiency and precision of pipeline laying. Compared with traditional methods, the present system can improve the pipeline laying efficiency by about 300%, and at the same time control the installation and assembly within millimeter level, which is of great significance for precise assembly work in the aircraft manufacturing process.

[0138] The present application adopts the fusion technology of high-precision laser radar and inertial measurement unit, cooperates with the innovative coordinate conversion algorithm, realizes the high-precision three-dimensional reconstruction of the complex aviation cabin environment. This not only provides reliable spatial information for subsequent path planning, but also greatly improves the adaptability of the system to complex environments. Especially in the narrow and winding aircraft cabin space, the system shows excellent spatial perception ability, laying a foundation for accurate layout of pipelines. The improved probability density estimation method used in the three-dimensional radiation field model construction of the present application significantly improves the calculation efficiency while ensuring the accuracy. The dynamic programming algorithm used in the clamp position calculation of the present application not only ensures the mechanical properties, but also minimizes the material usage, reflecting the intelligence and economy of the system.

[0139] The application innovatively combines graph neural networks and deep reinforcement learning techniques to develop an efficient and flexible path planning algorithm. This algorithm not only quickly generates optimal pipe laying paths, but also dynamically adjusts the planning results in real time in response to environmental changes. In particular, in complex scenarios with multiple parallel pipe laying, the system's interference avoidance capability is outstanding, effectively reducing the risk of interference between pipes and other structural components. BRIEF DESCRIPTION OF DRAWINGS

[0140] Figure 1 A principle diagram of an aircraft system pipe intelligent laying system of the application is shown in the figure;

[0141] Figure 2 A principle diagram of a deep vision subsystem is shown in the figure;

[0142] Figure 3 A principle diagram of an intelligent path planning subsystem is shown in the figure;

[0143] Figure 4 A principle diagram of an augmented reality guide subsystem is shown in the figure. DETAILED DESCRIPTION

[0144] Embodiment 1:

[0145] An aircraft system pipe intelligent laying system, as shown in the figure, comprises a deep vision subsystem, an intelligent path planning subsystem and an augmented reality guide subsystem connected in sequence from front to back, for realizing the intelligentization and high efficiency of aircraft system pipe laying. Figures 1-4

[0146] The deep vision subsystem is used to acquire three-dimensional data and positioning information of the aircraft system pipe, generate a three-dimensional model of the pipe, and determine the position information of the point cloud data in the world coordinate system according to the inertial measurement unit coordinate system.

[0147] The intelligent path planning subsystem is used to receive the three-dimensional model and point cloud position information sent by the deep vision subsystem, generate a high-precision three-dimensional radiation field model based on the three-dimensional model and point cloud position information, and plan an optimal pipe laying path in the high-precision three-dimensional radiation field model using a graph neural network and a reinforcement learning algorithm.

[0148] The augmented reality guide subsystem is used to receive the optimal pipe laying path sent by the intelligent path planning subsystem, detect the pipe position in real time through a laser sensor, superimpose the optimal pipe laying path on the real-time detected pipe position, and project the best installation position of the pipe clamp in real time through augmented reality glasses.

[0149] ​Preferably, the depth vision subsystem comprises a laser radar module, an inertial measurement unit module, an optical encoder, and a coordinate conversion module. The inertial measurement unit module is in communication connection with the laser radar module, and the optical encoder is in communication connection with the laser radar module and the inertial measurement unit module respectively; and the coordinate conversion module is in communication connection with the inertial measurement unit module.

[0150] The laser radar module is used to acquire three-dimensional data and positioning information of the aircraft system pipeline.

[0151] The inertial measurement unit module is used to acquire spatial pose information of an onboard inertial sensor; and based on the spatial pose information, the position of the three-dimensional data of the aircraft system pipeline measured by the laser radar in the onboard inertial sensor coordinate system is calculated.

[0152] The optical encoder is used to acquire the scanning frequency of the laser radar module; and record the current timestamp of the inertial measurement unit module.

[0153] The coordinate conversion module is used to receive the three-dimensional data in the onboard inertial sensor coordinate system sent by the inertial measurement unit module; and convert the three-dimensional data in the onboard inertial sensor coordinate system into point cloud position information in the world coordinate system.

[0154] Specifically, the laser radar module adopts a high-precision phase modulation continuous wave (FMCW) laser radar, which has a ranging accuracy of millimeter level and a scanning frequency of up to 20 Hz, so that the system can quickly and accurately acquire three-dimensional data of the aircraft system pipeline.

[0155] The inertial measurement unit module adopts micro-electro-mechanical system (MEMS) technology, integrates a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, and can provide six-degree-of-freedom attitude information. The sampling rate of the module is as high as 1000 Hz, ensuring the real-time and accuracy of the spatial pose information.

[0156] The resolution of the optical encoder reaches 0.1°, which can accurately capture the change of the scanning frequency of the laser radar module. This high-precision time synchronization mechanism is crucial for subsequent data fusion.

[0157] Preferably, the intelligent path planning subsystem comprises a radiation field generation module, a graph neural network module, a reinforcement learning module, and a pipeline interference avoidance module; the graph neural network module is in communication connection with the radiation field generation module, the reinforcement learning module is in communication connection with the graph neural network module, and the pipeline interference avoidance module is in communication connection with the reinforcement learning module.

[0158] a radiation field generation module, configured to receive the three-dimensional model and the point cloud position information sent by the depth vision subsystem, divide a three-dimensional space into a plurality of grids, calculate a probability density of each grid, and generate a high-precision three-dimensional radiation field model based on the probability density;

[0159] a graph neural network module, configured to receive the high-precision three-dimensional radiation field model sent by the radiation field generation module, extract a feature vector of each grid, and calculate a probability of each grid as a pipeline end point based on the feature vector;

[0160] a reinforcement learning module, configured to receive the grid probability information sent by the graph neural network module, select an action policy based on an ε-greedy algorithm, and generate an optimal pipeline laying path through iterative optimization;

[0161] a pipeline interference avoidance module, configured to receive the optimal pipeline laying path sent by the reinforcement learning module, define a pipeline interference constraint condition, and optimize the optimal pipeline laying path based on the interference constraint condition.

[0162] Specifically, the radiation field generation module adopts an improved fast Fourier transform (FFT) algorithm, divides the space into 64x64x64 grids, and the edge length of each grid is 5 cm. This division method greatly improves the calculation efficiency while ensuring the accuracy.

[0163] The graph neural network module adopts a 5-layer graph convolution network structure, and the number of neurons in each layer is 64, 128, 256, 128, and 64 respectively. In the training process, the Adam optimizer is adopted, the learning rate is set to 0.001, and the batch size is 32. After a large number of experiments, these hyperparameter settings can achieve the best performance in the aircraft system pipeline laying task.

[0164] Preferably, the augmented reality guidance subsystem includes a laser sensor module, a simulation rendering module, a clamp position calculation module, and an AR projection module; the simulation rendering module is in communication connection with the laser sensor module, the clamp position calculation module is in communication connection with the simulation rendering module, and the AR projection module is in communication connection with the simulation rendering module and the clamp position calculation module respectively. These modules work cooperatively to provide intuitive and accurate visual guidance for the actual installation process of the aircraft system pipeline.

[0165] The laser sensor module is configured to detect the pipeline position in real time and generate pipeline position data.

[0166] The simulation rendering module is configured to receive the optimal pipeline laying path sent by the intelligent path planning subsystem, superimpose the optimal pipeline laying path on the pipeline position data, and generate an augmented reality scene.

[0167] The clamp position calculation module is configured to determine the clamp positions in the optimal pipeline laying path according to a preset distance threshold, and generate clamp position data.

[0168] The AR projection module is configured to receive the augmented reality scene and the clamp position data, and project the optimal installation positions of the pipeline clamps in real time through AR glasses.

[0169] Specifically, the laser sensor module adopts a high-precision structured light laser scanning technology, which can capture the spatial position and attitude of the pipeline in real time. The scanning frequency of the module can reach 30 Hz, and the measurement accuracy is better than 0.1 mm. Such high temporal and spatial resolution characteristics enable the system to respond to small changes in the position of the pipeline in a timely manner, providing a reliable data foundation for subsequent real-time rendering and guidance.

[0170] The simulation rendering module receives the optimal pipeline laying path from the intelligent path planning subsystem and superimposes it on the pipeline position data detected by the laser sensor module in real time. This module adopts an innovative adaptive rendering algorithm that can dynamically adjust rendering parameters according to environmental lighting conditions and pipeline material properties.

[0171] The AR projection module is the key interface between the system and the operator. This module adopts multiple contour line simulation technology to present the clamp installation positions in an intuitive visual form in AR glasses. This contour-based AR projection method has the following advantages compared to traditional single-point marking methods: first, it provides continuous spatial information, enabling the operator to more intuitively understand the installation position of the clamp; second, the multi-level contour line can provide depth information, helping the operator accurately determine the spatial position of the installation point; finally, this method has relatively low requirements for the positioning accuracy of AR glasses, and has better robustness.

[0172] The augmented reality guidance subsystem of the present application can provide precise and intuitive installation guidance for the operator. In practical applications, this intelligent guidance system can improve the efficiency of pipeline installation by more than 50%, while controlling the installation error within 1 mm, significantly improving the quality and efficiency of aircraft system pipeline laying.

[0173] Through the cooperative work of the above various modules, the pipeline laying task can be quickly and accurately completed, and the work efficiency and installation quality are greatly improved. Compared with the traditional method, the system improves the laying speed by about 300%, and the installation accuracy is improved by about 50%, greatly reducing the assembly error and work intensity of manual operation. The application can quickly and accurately plan the optimal pipeline laying path in the complex aircraft cabin environment, while effectively avoiding the interference between the pipelines. This not only greatly improves the efficiency of pipeline laying, but also significantly reduces the risk of human operation errors, providing an intelligent and automated solution for pipeline laying work in the aircraft manufacturing process.

[0174] Embodiment 2:

[0175] An aircraft system pipeline intelligent laying method, as shown in Figure 1 , comprising the following steps:

[0176] 1. First, based on the depth vision subsystem, the three-dimensional data and positioning information of the aircraft system pipeline are obtained.

[0177] Preferably, as shown in Figure 2 , a laser radar module is used to scan the environment at a frequency of 20Hz, and an inertial measurement unit module is used to record the spatial pose information at a sampling rate of 1000Hz. The scanning frequency of the laser radar is accurately captured by using an optical encoder, realizing high-precision time synchronization. Then, the coordinate conversion module converts the data in the airborne inertial sensor coordinate system into point cloud position information in the world coordinate system based on the Kalman filter algorithm, and the conversion error is controlled within 1cm.

[0178] Preferably, the coordinate conversion module adopts an improved Kalman filter algorithm, and its state equation and observation equation are as follows:

[0179] X k =F k X k-1 +B k U k +w k ;

[0180] Z k =H k X k +v k ;

[0181] Wherein, X k is the state vector, F k is the state transition matrix, B k is the control input matrix, U k is the control vector, w k is the process noise, Z k is the observation vector, and Hk is the observation matrix, v k is the observation noise. By adjusting the noise covariance matrix Q k and R k , the present application realizes high-precision conversion of the on-board inertial sensor coordinate system to the world coordinate system, with conversion error controlled within 1cm.

[0182] 2. Then, based on the intelligent path planning subsystem, the optimal pipeline laying path is generated.

[0183] (1) First, based on the radiation field generation module, a high-precision three-dimensional radiation field model is constructed, and the space is divided into a 64x64x64 grid.

[0184] Preferably, the present application adopts an improved radiation field generation algorithm, the mathematical expression of which is as follows:

[0185]

[0186] where F(x, y, z) represents the radiation field intensity at the spatial point (x, y, z);

[0187] (x i , y i , z i ) is the coordinate of the i-th sampling point;

[0188] N is the number of sampling points;

[0189] w i is the weight of the i-th sampling point;

[0190] φ() is the radial basis function.

[0191] The radiation field generation algorithm plays a crucial role in the present application, and its core is to construct a high-precision three-dimensional space representation as the basis for subsequent intelligent path planning. There is a clear logical relationship between this algorithm and the three-dimensional model and the probability density of the network, and the specific content is as follows:

[0192] First, the point cloud data obtained from the depth vision subsystem provides initial three-dimensional spatial information, and these point cloud data are converted into position information in the world coordinate system. Based on these spatial points, the radiation field generation algorithm constructs a continuous radiation field intensity F(x, y, z); the radiation field intensity here is essentially a weighted representation of the spatial point distribution, where each sampling point contributes its own radiation influence, which is spread through radial basis functions (RBF). The radial basis function uses a Gaussian kernel:

[0193]

[0194] where r is the linear distance between the target point and the sampling point; β is an adjustable parameter, usually ranging from 0.1 to 10. In an embodiment of the present application, β is set to 1.5, which is obtained through a large number of experiments and can achieve a good balance between calculation efficiency and accuracy.

[0195] Next, the radiation field algorithm discretizes the three-dimensional space into a regular grid (64x64x64) and calculates the probability density of each point belonging to each grid. Preferably, the radiation field generation module uses an improved probability density estimation method to generate a high-precision three-dimensional radiation field model. The core idea of this method is to divide the three-dimensional space into a grid and then calculate the probability density of each point belonging to each grid. Specifically, the probability density function of each point belonging to a certain grid is defined as follows:

[0196]

[0197] where x represents a point in the point cloud, c i represents the center of grid i, and σ is an adjustable parameter. In an embodiment of the present application, σ usually ranges from 0.1 to 1, and the specific value can be adjusted according to the density and distribution characteristics of the point cloud. For example, when the point cloud is sparse, a larger σ value (such as 0.8) can be selected to increase the smoothing effect; when the point cloud is dense and rich in details, a smaller σ value (such as 0.2) can be selected to retain more detailed information.

[0198] This probability density function establishes a soft correspondence between the point cloud data and the discrete grid, avoiding the information loss that may be caused by hard assignment. Subsequently, by combining the gray value (or pixel value) of the point, the feature vector of each grid is calculated:

[0199]

[0200] G(x, i) = I(x) · P(x | i);

[0201] where G(x, i) is the gray value of each point in the point cloud belonging to grid i;

[0202] I(x) represents the pixel value of point x in the image. This method combines the color information of the point with its spatial distribution, which can better capture the geometric and visual features of the environment;

[0203] F(i) represents the feature vector corresponding to grid i. This feature extraction method not only considers the spatial distribution of points, but also integrates color information, which can more comprehensively describe the environmental features and provide more reliable input for subsequent path planning.

[0204] This design makes the radiation field contain both spatial geometric information and visual features, providing rich input features for graph neural networks. The grid feature vector of the radiation field model becomes the initial feature of the graph neural network nodes, thereby establishing a bridge between the radiation field and the graph neural network. This conversion process from continuous space to discrete grid and then to graph structure enables efficient processing and optimization of complex three-dimensional environments.

[0205] (2) Then, based on the graph neural network module, the grid features are extracted and the terminal probability is calculated. Specifically, a 5-layer graph convolutional network structure is adopted, and the number of neurons at each layer is 64, 128, 256, 128, and 64, respectively.

[0206] Preferably, the graph neural network and reinforcement learning algorithm are also used to plan the optimal pipeline laying path in a high-precision three-dimensional radiation field model. The improved graph neural network structure adopted by the present application has the following core formula:

[0207]

[0208] wherein, represents the feature representation of node v at the kth layer;

[0209] σ() is the activation function;

[0210] W (k) and b (k) are learnable parameters;

[0211] is the feature representation of node u at the (k-1)th layer. Here, "node" corresponds to a grid cell in three-dimensional space. In the context of pipeline planning, each node represents a position in space, and its feature vector encodes the geometric characteristics of that position, the surrounding environment information, and the suitability of being part of a pipeline path.

[0212] is the neighborhood set of node v, i.e., all nodes directly connected to node v. In a three-dimensional grid, this usually refers to adjacent grid cells, corresponding to possible local movement directions for the pipeline.

[0213] AGGREGATE (k) is the aggregation function. In the present application, we choose the attention mechanism as the aggregation function to better capture the relationships between nodes:

[0214]

[0215] wherein a() is the attention score function, and the present application adopts the dot product attention mechanism.

[0216] (3) Based on the reinforcement learning module, the deep deterministic policy gradient (DDPG) algorithm is used for path planning, in which the Actor network and Critic network both adopt a three-layer fully connected structure with 256 and 128 hidden layer neurons.

[0217] Preferably, the reinforcement learning module uses the Double DQN algorithm, whose Q value update formula is as follows:

[0218] Q(s t , a t )←Q(s t , a t )+α[r t +γQ(s t+1 , argmax a Q(s t+1 , a))-Q(s t , a t )];

[0219] Where s t is the current state, a is the learning rate, γ is the discount factor, and r is the immediate reward. In the present application, a is set to 0.01 and γ is set to 0.99, which are the parameter values that have been repeatedly optimized to achieve fast convergence and stable performance in the aircraft system pipeline laying task.

[0220] (4) Finally, based on the pipeline interference avoidance module, a distance field-based collision detection method is applied to ensure that the generated path does not interfere with other structures.

[0221] Preferably, the pipeline interference avoidance module uses an improved collision detection algorithm, the core idea of which is to simplify the pipeline into a series of spheres and then use the distance between the spheres to determine whether there is interference. Specifically, for two pipeline segments i and j, we calculate the minimum distance between them:

[0222]

[0223] Where S i and S j represent the sphere set of pipeline segments i and j, respectively. If d ij is less than a preset threshold (5mm in the present application), it is considered that there is interference and the path needs to be adjusted.

[0224] Preferably, in the implementation process of the present application, the pipeline interference avoidance module uses a distance field-based collision detection method. First, the pipeline surface is discretized into a series of point sets {p i}, and then the minimum distance from each point to the obstacle is calculated:

[0225]

[0226] where o represents the set of obstacle points. Based on these distance values, a continuous distance field function can be constructed:

[0227]

[0228] where w i (x) is the weight function, usually inverse distance weighting:

[0229]

[0230] With this distance field function, it is possible to quickly detect whether the pipeline collides with obstacles.

[0231] If D(x) < d threshold (in this invention, d threshold is set to 5mm), it is considered that there is a risk of collision, and the path needs to be adjusted.

[0232] This distance field-based method has the advantages of high computational efficiency and strong adaptability to complex shapes compared to traditional geometric collision detection algorithms, and is particularly suitable for interference detection tasks in complex environments such as aircraft systems pipelines.

[0233] 3、Finally, based on the augmented reality guidance subsystem, guide the pipeline laying operation.

[0234] Preferably, the laser sensor module is used to detect the pipeline position in real time at a frequency of 30Hz, with a measurement accuracy of better than 0.1mm. The optimal pipeline laying path is superimposed on the real-time detected pipeline position using the simulation rendering module, and the adaptive rendering algorithm is used to dynamically adjust the rendering parameters. The clamp position calculation module uses a dynamic programming algorithm to determine the optimal clamp installation position. For aluminum alloy pipelines with a diameter of 10mm, the maximum distance between adjacent clamps is set to 500mm. Based on the AR projection module, multiple contour line simulation techniques are used, usually 3 to 5 contour lines, corresponding to distance values of 10mm, 20mm, 30mm, etc., and the best installation position of the pipeline clamp is projected in real time through AR glasses.

[0235] Preferably, specifically, the core formula of the rendering algorithm is as follows:

[0236] I(x,y)=k a I a +k d (L·N)I d +k s (R·V) n I s ;

[0237] where I(x, y) represents the final rendering intensity of pixel (x, y), k a , k d and k s are the ambient light, diffuse reflection and specular reflection coefficients respectively, I a , I d and I s are the intensities of ambient light, diffuse reflection light and specular reflection light respectively, L is the light source direction vector, N is the surface normal vector, R is the reflected light direction vector, V is the line of sight direction vector, and n is the specular reflection index.

[0238] In practical application of the present application, these parameters are dynamically adjusted according to real-time collected environmental data. For example, in a dark environment, the system appropriately increases the value of k_a to improve the overall brightness; and on the surface of a high-reflectivity material pipe, the system increases the value of k_s to enhance the specular reflection effect. This adaptive rendering technique greatly improves the visual quality and recognition of the augmented reality image, enabling the operator to clearly identify the pipe laying path in various complex environments.

[0239] Preferably, the clamp position calculation module is a key innovation point of the present application. Based on the optimal pipe laying path, the module uses a dynamic programming algorithm to determine the optimal installation position of the clamp. The core idea of the algorithm is to minimize the number of clamps used under the premise of meeting the mechanical constraints. Specifically, for the distance D(i, j) between any two points i and j on the path, the calculation is as follows:

[0240]

[0241] Then, a state transition function f(i) is defined, which represents the minimum number of clamps required from the starting point to the ith point:

[0242]

[0243] where D max is the maximum allowed distance between clamps, which is usually determined according to factors such as the material, diameter and pressure bearing capacity of the pipe. In an embodiment of the present application, for an aluminum alloy pipe with a diameter of 10 mm, D max is set to 500 mm. This value is obtained through a large number of experiments and finite element analysis, which can ensure the stability of the pipe while minimizing the number of clamps used.

[0244] By solving the above state transition equation through dynamic programming, the optimal distribution of clamps on the entire path can be obtained. This method not only ensures the mechanical stability of the pipe installation, but also maximizes the reduction of material usage and installation workload, resulting in significant economic benefits.

[0245] Preferably, the contour generation process can be represented as:

[0246] L(k) = {(x, y, z) | f(x, y, z) = k};

[0247] where f(x, y, z) represents the distance function to the center of the clamp, and k is a different distance value. In practical applications, usually 3 to 5 contours are selected, corresponding to k values of 10mm, 20mm, 30mm, etc. This multi-level visualization method can provide accurate spatial positioning information for the operator, greatly improving the accuracy of clamp installation.

[0248] The process of projecting the contour into the AR glasses can be represented as:

[0249] P(L(k)) = T L(k);

[0250] where T is the projection transformation matrix, which takes into account the optical characteristics of the AR glasses, the head posture of the operator, and the geometric structure of the environment. In the present application, the T matrix is constantly updated through a real-time calibration algorithm, ensuring that the projected image is always accurately aligned with the actual environment.

[0251] Preferably, the graph neural network module comprises:

[0252] a node feature extraction unit for calculating the relationship features of each node with surrounding nodes; generating a node feature vector;

[0253] a edge feature extraction unit, in communication connection with the node feature extraction unit, for calculating the weight relationship between nodes; generating an edge feature vector;

[0254] a graph convolution layer, in communication connection with the node feature extraction unit and the edge feature extraction unit, for aggregating node features and edge features; updating node representation;

[0255] an output layer, in communication connection with the graph convolution layer, for calculating the probability of each node becoming a pipeline endpoint.

[0256] Preferably, the graph neural network module adopts an innovative node feature extraction method. This method not only considers the features of the node itself, but also introduces the relationship information of surrounding nodes. Specifically, for each node p, the relationship feature R(p, q) between it and the surrounding node q is calculated:

[0257]

[0258] where ||p-q||2 represents the Euclidean distance between nodes p and q, and where p and q denote the i-th neighbor of node p and q, respectively, K is the number of considered neighbors (in the present invention, K usually takes a value between 5 and 10), and G(p, q) is a Gaussian kernel function:

[0259]

[0260] where σ is a tunable parameter, usually taking a value between 0.1 and 1. This relationship feature design can capture the local structural information between nodes, which helps to improve the representation ability of the graph neural network.

[0261] In terms of edge feature extraction, the present invention calculates the weight relationship between nodes:

[0262]

[0263] where β is a tunable parameter, and N(p) represents the neighborhood set of node p. In practical applications, the value of β usually ranges from 0.5 to 2. A larger β value will make the weight distribution more concentrated, while a smaller β value will make the weight distribution more uniform. The graph convolution layer of the present invention uses an attention mechanism to aggregate node features and edge features:

[0264]

[0265] where, denotes the feature representation of node v at the k-th layer, σ() is an activation function (ReLU function is used in the present invention), W (k) and b (k) are learnable parameters, and α vu is the attention coefficient:

[0266]

[0267] where a is a learnable attention vector, and || represents the vector concatenation operation. This attention mechanism can adaptively adjust the importance of different neighbor nodes, thereby improving the representation ability and generalization performance of the model.

[0268] Preferably, the reinforcement learning module adopts the Deep Deterministic Policy Gradient (DDPG) algorithm. This algorithm includes four main components: Actor network, Critic network, experience replay buffer, and target network.

[0269] The Actor network is responsible for generating actions based on the current state. Its structure is a three-layer fully connected neural network, with 256 and 128 neurons in the hidden layers, respectively. The output layer uses a tanh activation function to limit the action within the range [-1, 1]. The update of the Actor network is achieved through the policy gradient:

[0270]

[0271] where, θ μ represents the parameters of the Actor network, Q(s, a|θ Q ) represents the output of the Critic network.

[0272] The Critic network is used to evaluate the action value function, and its structure is also a three-layer fully connected neural network, and the number of neurons in the hidden layer is 256 and 128 respectively. The update of the Critic network is realized by minimizing the temporal difference error:

[0273]

[0274] where, y i =r i +γQ'(s i+1 ,μ'(s i+1 |θ μ′ )|θ Q′ ) is the target Q value, and Q' and μ' represent the target Critic network and the target Actor network respectively.

[0275] The experience replay buffer is used to store the state transition samples (s t , a t , r t , s t+1 ), and the size of the buffer is set to 106. During the training process, a batch is randomly sampled from the buffer each time (the batch size is set to 64) for learning, and this method can break the correlation between samples and improve the stability of learning.

[0276] The target network is used to stabilize the learning process, and the parameter θ' gradually approaches the parameter θ of the online network through soft update:

[0277] θ'←τθ+(1-τ)θ';

[0278] Where, τ is the soft update coefficient, which is set to 0.001 in the present application. This soft update mechanism can effectively reduce the fluctuation of the target value and improve the stability of the training.

[0279] The present application can realize the intelligent and efficient laying of the aircraft system pipeline, compared with the traditional manual laying method, the installation efficiency is improved by more than 3 times, and the installation error is controlled in the millimeter level, which provides a comprehensive intelligent solution for the pipeline laying work in the aircraft manufacturing process.

[0280] The above is only the preferred embodiment of the present application, and does not limit the present application in any form, and any simple modification, equivalent change of the above embodiment according to the technical essence of the present application falls within the protection scope of the present application.

Claims

1. A method for intelligent laying of aircraft system pipelines, characterized in that, Includes the following steps: Step S1: Obtain the three-dimensional data and positioning information of the aircraft system piping; Step S11: Collect three-dimensional data of the aircraft system piping, and at the same time, acquire the spatial pose information of the airborne inertial sensors; Step S12: Based on spatial pose information, calculate the three-dimensional data of the aircraft system piping in the airborne inertial sensor coordinate system; Step S13: Convert the 3D data in the airborne inertial sensor coordinate system into point cloud position information in the world coordinate system; Step S2: Generate the optimal pipeline laying path; Step S21: Based on the point cloud location information, a continuous radiation field is generated using a radiation field generation algorithm to construct a three-dimensional radiation field model; the three-dimensional space is divided into multiple grids, and the probability density of each grid is calculated; Based on the probability density, the gray value G(x, i) of the point is combined to calculate the grid feature vector F(i), which is then input into the graph neural network. Step S22: Based on the graph neural network, extract the grid features and calculate the probability of the pipeline endpoint; Step S23: Use the deep deterministic gradient algorithm to perform path planning and generate the pipeline laying path; Step S24: Based on pipeline interference constraints, obtain the optimal pipeline laying path; Step S3: Real-time detection of pipeline location, superimposing the optimal pipeline laying path onto the real-time detected pipeline location, and real-time projection of the optimal installation position of the pipeline clamp.

2. The intelligent laying method for aircraft system piping according to claim 1, characterized in that, In step S13, the Kalman filter algorithm is used for coordinate transformation, where the state equation and observation equation are: X k =F k X k-1 +B k U k +w k ; Z k =H k X k +v k ; Where: X k It is a state vector; X k-1 Let k be the state vector of the system at time k-1; F k This is the state transition matrix; B k To control the input matrix; U k For control vectors; w k This is process noise; Z k For observation vectors; H k The observation matrix; v k To observe noise.

3. The intelligent laying method for aircraft system piping according to claim 1, characterized in that, In step S21, the radiation field generation algorithm is as follows: Where F(x, y, z) represents the radiation field intensity at the spatial point (x, y, z); (x i y i , z i () represents the coordinates of the i-th sampling point; N is the number of sampling points; w i The weight of the i-th sampling point; φ() is a radial basis function; Calculate the probability density for each grid cell: Where: x is a point in the point cloud. c i Indicates the center of grid i. c i Indicates the center of grid j. σ is an adjustable parameter; The grayscale value G(x, i) of the point is: G(x, i) = I(x)·P(x|i); The feature vector F(i) for each grid cell:

4. The intelligent laying method for aircraft system piping according to claim 1, characterized in that, Step S22 includes the following steps: Step A1: Calculate the relationship features between each node and its surrounding nodes, and generate node feature vectors; Where R(p, q) represents the relationship features between each node and its surrounding nodes; ||pq||2 represents the Euclidean distance between nodes p and q; and Let p and q represent the i-th neighboring nodes of nodes p and q, respectively; K is the number of neighboring nodes to consider. G(p, q) is the Gaussian kernel function: Step A2: Calculate the weight relationships between nodes and generate edge feature vectors; Where: W(p, q) represents the weight relationship between node p and node q; β is an adjustable parameter; N(p) represents the neighborhood set of node p; ||pv||2 is the Euclidean distance between node p and node v; Step A3: Aggregate node features and edge features, and update the node representation; Among them: W (k) and b( k) These are learnable parameters; Let u be the feature representation of node u at the k-th layer; Let v be the feature representation of node v at the (k+1)th layer; α vu Attention coefficient; σ() is the activation function; N(v) is the neighborhood set of node v; Step A4: Calculate the probability that each node will become the end point of the pipeline.

5. A method for intelligent laying of aircraft system pipelines according to claim 1 or 4, characterized in that, In step S22, the formula for the graph neural network is: in: This represents the feature representation of node v at the k-th layer. σ() is the activation function; W (k) and b (k) These are learnable parameters; Let u be the feature representation of node u at the (k-1)th layer; N(v) is the neighborhood set of node v; AGGREGATE (k) For aggregate functions: Where: a() is the attention scoring function; α vu Attention coefficient; u represents a node within the neighborhood; Let v be the feature representation of node v at the (k-1)th layer; Let be the feature vector of the neighbor node w in the kl-th layer of the graph neural network.

6. The intelligent laying method for aircraft system piping according to claim 1, characterized in that, Step S24 includes the following steps: Step B1: Discretize the pipe surface into a series of point sets {p i Then calculate the minimum distance from each point to the obstacle: Where o represents the set of obstacle points; Step B2: Construct a continuous range field function: Among them, w i (x) is the weighting function: Step B2: If D(x) < d threshold If so, it is considered that there is a risk of collision; Where: d threshold This is the preset distance threshold.

7. The intelligent laying method for aircraft system piping according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Use a laser sensor to detect the pipeline position in real time and generate pipeline position data; Step S32: Overlay the optimal pipeline laying path onto the pipeline location data, and use an adaptive rendering algorithm to dynamically adjust the rendering parameters to generate an augmented reality scene; The adaptive rendering algorithm is as follows: I(x,y)=k a I a +k d (L·N)I d +k s (R·V) n I s ; Where: I(x, y) represents the final rendering intensity of pixel (x, y); k a k d and k s These are the ambient light, diffuse reflection, and specular reflection coefficients, respectively. I a I d and I s These represent the intensities of ambient light, diffuse light, and specular light, respectively. L is the direction vector of the light source; N is the surface normal vector; R is the direction vector of the reflected ray; V is the line-of-sight vector; n is the specular reflection index; Step S33: Calculate the optimal installation position of clamps in the optimal pipeline laying path, where the number of clamps used is minimized while satisfying mechanical constraints; Define a state transition function f(i) to represent the minimum number of clamps required to get from the starting point to the i-th point: Among them, D max This refers to the maximum permissible distance between the clamps; Any point i(x) on the path D(i,j) i y i , z i ) and point j(x j y j , z j The distance between them; Step S34: Using multiple contour line simulation technology, the optimal installation position of the pipeline clamp is projected in real time through AR glasses.

8. An intelligent piping system for aircraft systems, used to implement the intelligent piping method for aircraft systems as described in any one of claims 1-7, characterized in that, It includes a depth vision subsystem, an intelligent path planning subsystem, and an augmented reality guidance subsystem, which are connected sequentially from front to back; The depth vision subsystem is used to acquire 3D data and positioning information of aircraft system pipelines and generate 3D models of the pipelines. Based on the coordinate system of the inertial measurement unit, determine the point cloud position information in the world coordinate system; The intelligent path planning subsystem is used to generate a three-dimensional radiation field model based on the three-dimensional model and point cloud location information; and to plan the optimal pipeline laying path in the three-dimensional radiation field model using graph neural networks and reinforcement learning algorithms. The augmented reality guidance subsystem is used to detect pipeline locations in real time and overlay the optimal pipeline laying path onto the detected pipeline locations; the optimal installation position of pipeline clamps is projected in real time through augmented reality glasses.

9. The intelligent piping laying system for aircraft systems according to claim 8, characterized in that, The depth vision subsystem includes a lidar module, an inertial measurement unit module, a coordinate transformation module, and an optical encoder connected sequentially from front to back. The optical encoder is communicatively connected to the lidar module and the inertial measurement unit module, respectively. The lidar module is used to acquire three-dimensional data and positioning information of the aircraft's system piping. The inertial measurement unit module is used to acquire the spatial pose information of the airborne inertial sensor and calculate the position of the three-dimensional data of the aircraft system piping in the airborne inertial sensor coordinate system. An optical encoder is used to acquire the scanning frequency of the lidar module and record the current timestamp of the inertial measurement unit module. The coordinate transformation module is used to convert 3D data in the airborne inertial sensor coordinate system into point cloud position information in the world coordinate system.

10. The intelligent piping laying system for aircraft systems according to claim 8, characterized in that, The intelligent path planning subsystem includes a radiation field generation module, a graph neural network module, a reinforcement learning module, and a pipeline interference avoidance module, which are connected sequentially from front to back. The radiation field generation module is used to divide the three-dimensional space into multiple grids based on the three-dimensional model and point cloud location information; calculate the probability density of each grid and generate a three-dimensional radiation field model. The graph neural network module is used to receive the three-dimensional radiation field model, extract the feature vector of each grid, and calculate the probability of each grid being the end point of the pipeline. The reinforcement learning module receives probability information from the grid and selects action strategies based on the ε-greedy algorithm; through iterative optimization, it generates pipeline laying paths. The pipeline interference avoidance module is used to receive the pipeline laying path and obtain the optimal pipeline laying path based on interference constraints.

11. The intelligent piping laying system for aircraft systems according to claim 8, characterized in that, The augmented reality guidance subsystem includes a laser sensor module, a simulation rendering module, a clamp position calculation module, and an AR projection module connected from front to back. The laser sensor module is used to detect the pipeline position in real time and generate pipeline position data; The simulation rendering module is used to receive the optimal pipeline laying path and overlay the optimal pipeline laying path onto the pipeline location data to generate an augmented reality scene. The clamp position calculation module is used to determine the clamp position in the optimal pipeline laying path based on a preset distance threshold and generate clamp position data. The AR projection module is used to receive clamp position data and project the optimal installation position of the pipe clamp in real time through AR glasses.

Images