Accurate positioning and spraying thickness control system in airplane paint spraying hangar

Abstract

The invention relates to the technical field of aircraft manufacturing, and discloses a precise positioning and spraying thickness control system in an aircraft paint spraying hangar, which consists of a data acquisition unit, a model construction unit and a spraying control unit. The data acquisition unit collects the spraying position, thickness and reference data in real time through a high-precision sensor, and a training data set is formed. The model construction unit constructs a positioning and thickness control model based on a machine learning algorithm, and performs correction and optimization through a residual error fitting model. And the spraying control unit generates a spraying position and thickness adjusting instruction according to the trained model in combination with real-time data, so that accurate spraying is realized. The sensor technology, the data processing technology and the machine learning technology are integrated, the automation and intelligence level of paint spraying operation is improved, the spraying uniformity and consistency are ensured, the overall quality of an airplane is improved, meanwhile, paint resources are saved, the manufacturing cost is reduced, and remarkable technical advantages and application value are achieved.

Description

Technical Field

[0001] The invention relates to the technical field of aircraft manufacturing, and in particular to a precise positioning and spraying thickness control system in an aircraft painting hangar. Background Art

[0002] In the process of aircraft manufacturing and maintenance, painting is an indispensable part. It is not only related to the appearance of the aircraft, but also directly affects the flight performance and safety performance of the aircraft. There are many shortcomings in the existing aircraft painting operations. On the one hand, due to the large size and complex structure of the aircraft, it is difficult to achieve high-precision requirements for the paint position, which often leads to problems such as uneven spraying, missed spraying or over-spraying, which not only wastes paint resources, but also may affect the overall quality of the aircraft. On the other hand, the control of spraying thickness is also crucial. Too thin may lead to insufficient protective performance of the coating, while too thick may increase the weight of the aircraft and affect flight efficiency. At present, there is a lack of effective thickness control methods, and it is difficult to ensure the uniformity and consistency of each layer of paint.

[0003] With the advancement of science and technology, especially the rapid development of sensor technology, data processing technology and machine learning technology, the automation and intelligence of aircraft painting operations have become possible. However, the integrated application of these technologies in the field of aircraft painting is still insufficient. The existing painting systems can often only achieve a single positioning or thickness control function, and the accuracy and stability need to be improved. Especially in a complex and changeable painting environment, how to obtain the spraying position data and spraying thickness data in real time and accurately, and build an efficient and accurate control model based on these data, has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The object of the present invention is to provide a precise positioning and spraying thickness control system in an aircraft painting hangar to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a precise positioning and spraying thickness control system in an aircraft painting hangar, the system comprising: The data acquisition unit is used to collect the spraying position data and spraying thickness data in real time during the aircraft painting process, which is achieved through the high-precision positioning sensor and thickness measurement sensor installed in the paint hangar; at the same time, the aircraft model information, paint process parameters and historical spraying records are collected as spraying reference data; the real-time data is integrated and pre-processed with the spraying reference data to form a spraying training data set; A model building unit is used to build a spraying position positioning model and a spraying thickness control model, wherein the positioning model is used to determine the precise position of the paint spraying equipment, and the thickness control model is used to adjust the thickness of the spray layer to meet a predetermined standard; based on the spraying training data set, a machine learning algorithm is used to train the positioning model and the thickness control model respectively, and the residuals of the two models are used to build a residual fitting model, and the residual fitting model is used to correct and optimize the results of positioning and thickness control; The spray control unit uses the trained model and combines it with real-time spray data to generate the aircraft's spray position adjustment instructions and spray thickness adjustment instructions.

[0006] Preferably, a convolutional neural network (CNN) algorithm is used to construct a spraying position positioning model, and high-dimensional features of the spraying position are extracted through multi-layer convolution and pooling operations.

[0007] Preferably, the training step of the spraying position positioning model includes: Construct the convolutional neural network (CNN) structure: Design the CNN model, determine the number of convolutional layers, convolution kernel size, pooling method, and structure of the fully connected layer; Model training: The spray training data set is input into the CNN model, and the model parameters are optimized by the back propagation algorithm and the gradient descent algorithm; the formula for optimizing the model parameters by the gradient descent algorithm is: ; in, Represents the model parameters, including all weights and biases that the model needs to learn; Represents the learning rate, which is used to control the step size of parameter update; Represents the loss function About parameters The gradient of , which is a vector pointing in the direction of fastest growth of the loss function; Indicates an assignment operation, that is, updating parameters The value of Model verification: Evaluate the positioning accuracy of the model on the verification set, and adjust the number of layers and convolution kernel size of the CNN model based on the evaluation results.

[0008] Preferably, a long short-term memory network (LSTM) algorithm is used to construct a spraying thickness control model, and the variation trend of the spraying thickness is predicted through time series analysis.

[0009] Preferably, the training step of the spraying thickness control model includes: Data serialization processing: Arrange the thickness data in the spraying training data set in time series; Construct the LSTM structure of the long short-term memory network: define the LSTM model, determine the number of hidden layers, the number of hidden units, and the learning rate; Model training: Input serialized data into the LSTM model and predict the change of spraying thickness through the learning ability of memory units; Model optimization: According to the error between the predicted result and the actual thickness, the parameters and structure of the LSTM model are adjusted.

[0010] Preferably, the step of constructing the residual fitting model includes: Calculate the residuals between the actual outputs of the spraying position positioning model and the spraying thickness control model and the true values, which are recorded as positioning residuals and thickness residuals respectively; The positioning residual and the thickness residual are combined with the original spraying data to form a residual feature set; The neural network algorithm is used to train the residual feature set and construct the residual fitting model; The structure of the residual fitting model includes an input layer, a hidden layer and an output layer. The input layer receives the residual feature set, and the output layer outputs the corrected spraying position and thickness control results, and reduces the residual through the nonlinear fitting ability of the neural network.

[0011] Preferably, the data acquisition unit preprocesses the collected data by means of data denoising and normalization.

[0012] Preferably, the method of denoising the data includes: A filtering algorithm is used to remove high-frequency noise from the data collected by the positioning sensor and the thickness measurement sensor; For abnormal data points caused by equipment failure, the average value of adjacent data points is used for replacement.

[0013] Preferably, the method of normalizing the data includes: performing linear transformation on the positioning data and thickness data according to their maximum and minimum values ​​so that their value ranges are mapped to the interval [0, 1].

[0014] Compared with the prior art, the present invention has the following beneficial effects: By integrating high-precision positioning sensors, this system can capture the specific position of the paint spraying equipment on the aircraft surface in real time and accurately, effectively solving the positioning problem caused by the large size and complex structure of the aircraft in traditional painting operations. This ensures the accuracy of the spraying operation and avoids problems such as uneven spraying, missed spraying or over-spraying, thereby greatly improving the quality and efficiency of the painting operation. The system's built-in thickness measurement sensor can monitor the thickness of the spray layer in real time, and combine with machine learning algorithms to build an accurate spray thickness control model. This innovative design makes the control of spray thickness more intelligent and automated, and can accurately control the thickness of each layer of paint, ensuring the protective performance of the coating while avoiding the problem of increasing the weight of the aircraft due to excessively thick coatings, which is of great significance to improving the flight efficiency of the aircraft.

[0015] This system not only collects spraying position data and thickness data in real time, but also integrates multi-dimensional data such as aircraft model information, painting process parameters, and historical spraying records to form a rich spraying training data set. The positioning model and thickness control model are trained through machine learning algorithms, and the residual fitting model is used for correction and optimization, which further improves the accuracy and stability of the system and ensures the consistency and reliability of the spraying operation. The application of this system has significantly improved the automation and intelligence level of the painting operation, reduced manual intervention and repetitive labor, and thus greatly improved the operation efficiency. At the same time, due to the precise control of the spraying quality and thickness, the waste of paint and rework caused by quality problems are reduced, effectively reducing the cost of the painting operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a working principle diagram of the precise positioning and spraying thickness control system in the aircraft painting hangar of the present invention; Figure 2 Flowchart for building a spraying location positioning model; Figure 3 Training flowchart for fitting a model to residuals. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] See also Figure 1-3 The present invention provides a technical solution: a precise positioning and spraying thickness control system in an aircraft painting hangar, the system comprising: Data acquisition unit: The data acquisition unit is responsible for collecting spraying position data and spraying thickness data in real time during the aircraft painting process. To achieve this function, this embodiment installs high-precision positioning sensors and thickness measurement sensors in the paint hangar. The high-precision positioning sensor uses laser ranging, visual positioning or RFID technology to accurately obtain the position information of the paint spraying equipment relative to the aircraft surface. The thickness measurement sensor uses ultrasonic, electromagnetic induction or optical principles to measure the thickness of the coating in real time.

[0019] The data acquisition unit is also responsible for collecting aircraft model information, painting process parameters (such as paint type, spraying pressure, spraying speed, etc.) and historical spraying records. This information is used as spraying reference data, and together with the real-time collected position data and thickness data, it is integrated and pre-processed to form a spraying training data set.

[0020] Model building unit: The model building unit builds a spraying position positioning model and a spraying thickness control model based on the spraying training data set. The positioning model uses machine learning algorithms (such as support vector machines, neural networks, etc.) to learn the mapping relationship between the spraying position and sensor data through training, so as to accurately determine the precise position of the spraying equipment. The thickness control model adjusts the thickness of the spray layer to meet the predetermined standard by learning the relationship between the spraying thickness and the spraying parameters.

[0021] In order to improve the accuracy and stability of the model, the present embodiment also uses the residuals of the positioning model and the thickness control model to construct a residual fitting model. The residual fitting model further analyzes and learns the residuals, corrects and optimizes the results of positioning and thickness control, and ensures the accuracy and consistency of the spraying operation.

[0022] Spraying control unit: The spraying control unit generates spraying position adjustment instructions and spraying thickness adjustment instructions for the aircraft based on the trained model and real-time spraying data. Specifically, the spraying control unit receives the real-time position data and thickness data from the data acquisition unit, inputs them into the positioning model and thickness control model, and obtains the deviation of the spraying position and the deviation of the thickness. Then, according to these deviations, corresponding adjustment instructions are generated to control the position and spraying parameters of the paint spraying equipment to achieve precise spraying.

[0023] The present invention will be further described below in conjunction with Examples 1 to 3: Embodiment 1:

[0024] This embodiment provides a method for constructing a spraying position positioning model based on a convolutional neural network (CNN) algorithm, which aims to extract high-dimensional features of the spraying position through multi-layer convolution and pooling operations to achieve accurate positioning in aircraft painting operations. The specific steps include: ① Construct the convolutional neural network CNN structure: Design CNN model: According to the characteristics of the spraying position positioning task, the structure of the CNN model is designed. The model includes an input layer, multiple convolutional layers, a pooling layer, a fully connected layer, and an output layer. The input layer receives the image data in the spraying training data set, the convolutional layer extracts the features of the image through convolution operations, the pooling layer performs dimensionality reduction on the features, the fully connected layer maps the high-dimensional features to the output space, and the output layer gives the prediction results of the spraying position.

[0025] Determine the number of convolutional layers, convolution kernel size, pooling method, and structure of the fully connected layer: Based on experimental experience and task requirements, the number of convolutional layers is set to N (N is a positive integer), and the convolution kernel size is selected according to the scale of the image features. Usually a smaller convolution kernel (such as 3x3 or 5x5) is selected to capture local features. The pooling method uses maximum pooling or average pooling to reduce the size of the feature map and retain important information. The structure of the fully connected layer is designed according to the dimension of the feature map and the dimension of the output space to ensure that the features can be mapped to the spraying position.

[0026] ② Perform model training: Input spray training dataset: Input the preprocessed spray training dataset into the CNN model. The dataset includes image data and corresponding spray position labels, which are used to train the model to learn the mapping relationship between image features and spray positions.

[0027] Optimize model parameters through back propagation algorithm and gradient descent algorithm: During the training process, the back propagation algorithm is used to calculate the gradient of the loss function with respect to the model parameters, and then the gradient descent algorithm is used to update the model parameters. The formula for optimizing the model parameters by the gradient descent algorithm is: ; in, Represents the model parameters, including all weights and biases that the model needs to learn; Represents the learning rate, which is used to control the step size of parameter update; Represents the loss function About parameters The gradient of , which is a vector pointing in the direction of fastest growth of the loss function; Indicates an assignment operation, that is, updating parameters By continuously iteratively updating the parameters, the loss function is gradually reduced and the model performance is improved.

[0028] Setting the learning rate and loss function: Learning rate The choice of is crucial to model training. Too large a learning rate may cause the model to fail to converge, while too small a learning rate will slow down the training process. This implementation method determines the appropriate learning rate through experiments. Loss function Common functions such as mean squared error (MSE) or cross entropy loss are used to measure the difference between the model's predicted results and the actual results.

[0029] ③ Model verification: Evaluate the positioning accuracy of the model on the validation set: Test the trained CNN model on the validation set to evaluate the positioning accuracy of the model. The validation set is a data set independent of the training set and is used to verify the generalization ability of the model.

[0030] Adjust the number of layers and convolution kernel size of the CNN model according to the evaluation results: According to the evaluation results on the validation set, if the model positioning accuracy is not ideal, the number of layers of the CNN model (increase or decrease convolution layers, pooling layers, etc.) and the size of the convolution kernel can be adjusted to further improve the model performance. Through continuous iterative training, verification and adjustment of the model structure, a spraying position positioning model with high-precision positioning capabilities is finally obtained.

[0031] Embodiment 2: This embodiment provides a method for constructing a spraying thickness control model based on a long short-term memory network (LSTM) algorithm. The method uses time series analysis to predict the change trend of the spraying thickness to achieve precise control of the spraying thickness. The specific steps of the method include: ① Data serialization processing: Extract thickness data from the spraying training dataset: Extract data related to spraying thickness from the spraying training dataset. These data should contain spraying thickness measurements at different time points.

[0032] Arrange thickness data in time series: Arrange the extracted thickness data in time order to form a time series data. This sequence data will be used as the input of the LSTM model to learn the change trend of the spraying thickness.

[0033] ② Constructing the long short-term memory network LSTM structure: Define the LSTM model: According to the requirements of the spraying thickness control task, define the structure of the LSTM model. The LSTM model includes an input layer, a hidden layer, and an output layer, where the hidden layer consists of multiple LSTM memory units.

[0034] Determine the number of hidden layers, the number of hidden units, and the learning rate: The number of hidden layers, the number of hidden units in each hidden layer, and the learning rate of the LSTM model are determined through experiments and experience. The choice of these parameters has an important impact on the performance and training speed of the model.

[0035] Initialize model parameters: Before model training begins, initialize the parameters of the LSTM model, including weights and biases.

[0036] ③ Model training: Input serialized data into the LSTM model: Input the processed time series data into the LSTM model as the training data of the model.

[0037] Predicting changes in spray thickness through the learning ability of memory units: The LSTM model uses the learning ability of its memory units to learn the input time series data to predict the trend of changes in spray thickness. During the training process, the model will continuously update its internal state based on the input data and historical information to improve the accuracy of the prediction.

[0038] Calculate prediction error and update model parameters: After each prediction, calculate the error between the prediction result and the actual thickness, and pass the error back to the model through the back propagation algorithm to update the model parameters. Through continuous iterative training, the model parameters are gradually optimized and the prediction error is gradually reduced.

[0039] ④ Model optimization: Evaluate model performance: During the training process, regularly evaluate the performance of the model on the validation set, including prediction accuracy, stability, and generalization ability.

[0040] Adjust the model parameters and structure according to the error between the predicted result and the actual thickness: According to the evaluation results on the validation set, if the prediction error is large or the model performance does not meet the requirements, the parameters and structure of the LSTM model need to be adjusted. For example, the model can be optimized by increasing the number of hidden layers, adjusting the number of hidden units, changing the learning rate, or introducing regularization techniques.

[0041] Iterative training and optimization: Through continuous iterative training and optimization of the model, until satisfactory prediction accuracy and performance requirements are achieved, the resulting LSTM model can accurately predict the trend of spray thickness changes and provide strong support for spray thickness control.

[0042] Embodiment 3: This embodiment describes in detail a method for constructing a residual fitting model, which aims to further improve the accuracy and efficiency of the spraying operation by calculating the residuals between the actual outputs of the spraying position positioning model and the spraying thickness control model and the true values, and fitting these residuals using a neural network algorithm. The steps of the method include: ① Calculate the residual: Obtain the actual output of the spray position positioning model and the spray thickness control model: During the spraying operation, the output data of the spray position positioning model and the spray thickness control model are collected in real time. These data represent the spray position and spray thickness predicted by the model, respectively.

[0043] Get the true value of spraying position and thickness: Get the true value of spraying position and thickness through high-precision measuring equipment or manual inspection as a reference standard for model output.

[0044] Calculate positioning residuals and thickness residuals: Compare the actual output of the spray position positioning model with the true value to calculate the positioning residual; similarly, compare the actual output of the spray thickness control model with the true value to calculate the thickness residual. These residuals reflect the difference between the model prediction and the true value.

[0045] ② Forming residual feature set: Extract original spraying data: Extract original spraying data from the spraying operation process, including spraying position, spraying thickness, spraying speed, spraying material and other related parameters.

[0046] Combining positioning residuals and thickness residuals: The calculated positioning residuals and thickness residuals are combined with the original spraying data to form a feature set containing residual information. This feature set not only contains the basic parameters of the spraying operation, but also contains information on the model prediction error, providing a rich data basis for subsequent residual fitting.

[0047] ③ Construct residual fitting model: Design the neural network structure: Design the structure of the neural network according to the requirements of the residual fitting task. The neural network includes an input layer, a hidden layer, and an output layer. The input layer is responsible for receiving the residual feature set, the hidden layer extracts features through nonlinear transformation, and the output layer outputs the corrected spray position and thickness control results.

[0048] Determine the parameters of the neural network: Based on experimental experience and task requirements, determine the parameters of the neural network, including the number of hidden layers, the number of neurons in each layer, the selection of activation functions, the setting of learning rates, etc. The selection of these parameters has an important impact on the performance and training speed of the neural network.

[0049] Training the neural network: Input the residual feature set into the neural network, and optimize the parameters of the neural network through the back propagation algorithm and the gradient descent algorithm. During the training process, the neural network gradually learns the mapping relationship between the residual feature set and the corrected spray position and thickness control results.

[0050] Verify and test the neural network: After training is completed, the neural network is verified using the validation set to evaluate its performance. If the performance meets the requirements, the neural network is tested using the test set to further verify its generalization ability.

[0051] ④ Apply residual fitting model: Real-time correction of spray position and thickness: During the spraying operation, the output of the spray position positioning model and the spray thickness control model is obtained in real time, and the residual is calculated. Then the residual and the original spraying data are input into the residual fitting model to obtain the corrected spray position and thickness control results.

[0052] Improve spraying accuracy and efficiency: By correcting the residual fitting model, the prediction error of the spraying position positioning model and the spraying thickness control model can be reduced, thereby improving the accuracy and efficiency of the spraying operation. This is of great significance for improving the quality of aircraft painting operations, reducing material waste and reducing production costs.

[0053] The method for preprocessing the collected data by the data acquisition unit includes two main steps: data denoising and normalization processing.

[0054] Data denoising includes: Using filtering algorithms to remove high-frequency noise: During the spraying process, the positioning sensor and the thickness measurement sensor will collect raw data containing high-frequency noise. In order to remove this noise, this embodiment uses a filtering algorithm, such as Kalman filtering, mean filtering or median filtering.

[0055] Taking Kalman filtering as an example, the algorithm recursively estimates the state of the system and uses the measurement data to make corrections, thereby effectively removing high-frequency noise from the data. For each set of data collected by the positioning sensor and the thickness measurement sensor, the Kalman filter algorithm is applied to process it to obtain the denoised data.

[0056] Replace abnormal data points: During the data collection process, some abnormal data points may be generated due to equipment failure or external interference. These abnormal data points will have an adverse effect on subsequent data analysis and model training. In order to deal with these abnormal data points, this embodiment adopts a method of replacing the average value of adjacent data points. Specifically, for each abnormal data point, the average value of several adjacent data points (such as 3 or 5) before and after it is calculated, and the abnormal data point is replaced with the average value. This ensures the continuity and smoothness of the data.

[0057] Data normalization processing includes: Linear transformation normalization: In order to make the positioning data and thickness data have a unified dimension and value range, and to facilitate subsequent data analysis and model training, this embodiment adopts a linear transformation normalization method. Specifically, for each set of positioning data or thickness data, first find its maximum and minimum values. Then, use the linear transformation formula to map the data to the [0,1] interval.

[0058] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0059] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various 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. Precision positioning and spray thickness control system in aircraft paint hangar, characterized by: The system comprises: The data acquisition unit is used to collect the spraying position data and spraying thickness data in real time during the aircraft painting process, which is achieved through the high-precision positioning sensor and thickness measurement sensor installed in the paint hangar; at the same time, the aircraft model information, paint process parameters and historical spraying records are collected as spraying reference data; the real-time data is integrated and pre-processed with the spraying reference data to form a spraying training data set; A model building unit is used to build a spraying position positioning model and a spraying thickness control model, wherein the positioning model is used to determine the precise position of the paint spraying equipment, and the thickness control model is used to adjust the thickness of the spray layer to meet a predetermined standard; based on the spraying training data set, a machine learning algorithm is used to train the positioning model and the thickness control model respectively, and the residuals of the two models are used to build a residual fitting model, and the residual fitting model is used to correct and optimize the results of positioning and thickness control; The spray control unit uses the trained model and combines it with real-time spray data to generate the aircraft's spray position adjustment instructions and spray thickness adjustment instructions.

2. The precise positioning and spraying thickness control system in the aircraft painting hangar according to claim 1 is characterized in that: The convolutional neural network (CNN) algorithm is used to build a spraying position positioning model, and the high-dimensional features of the spraying position are extracted through multi-layer convolution and pooling operations.

3. The precise positioning and spraying thickness control system in the aircraft painting hangar according to claim 2 is characterized in that: The training steps of the spraying position positioning model include: Construct the convolutional neural network (CNN) structure: Design the CNN model, determine the number of convolutional layers, convolution kernel size, pooling method, and structure of the fully connected layer; Model training: The spray training data set is input into the CNN model, and the model parameters are optimized by the back propagation algorithm and the gradient descent algorithm; the formula for optimizing the model parameters by the gradient descent algorithm is: ； in, Represents the model parameters, including all weights and biases that the model needs to learn; Represents the learning rate, which is used to control the step size of parameter update; Represents the loss function About parameters The gradient of , which is a vector pointing in the direction of fastest growth of the loss function; Indicates an assignment operation, that is, updating parameters The value of Model verification: Evaluate the positioning accuracy of the model on the verification set, and adjust the number of layers and convolution kernel size of the CNN model based on the evaluation results.

4. The precise positioning and spraying thickness control system in the aircraft painting hangar according to claim 1 is characterized by: The long short-term memory network (LSTM) algorithm is used to build a spraying thickness control model, and the changing trend of spraying thickness is predicted through time series analysis.

5. The precise positioning and spraying thickness control system in the aircraft painting hangar according to claim 4 is characterized in that: The training steps of the spraying thickness control model include: Data serialization processing: Arrange the thickness data in the spraying training data set in time series; Construct the LSTM structure of the long short-term memory network: define the LSTM model, determine the number of hidden layers, the number of hidden units, and the learning rate; Model training: Input serialized data into the LSTM model and predict the change of spraying thickness through the learning ability of memory units; Model optimization: According to the error between the predicted result and the actual thickness, the parameters and structure of the LSTM model are adjusted.

6. The precise positioning and spraying thickness control system in an aircraft painting hangar according to claim 1, characterized in that: The steps to construct a residual fitting model include: Calculate the residuals between the actual outputs and true values ​​of the spraying position positioning model and the spraying thickness control model, which are recorded as positioning residuals and thickness residuals respectively; The positioning residual and the thickness residual are combined with the original spraying data to form a residual feature set; The neural network algorithm is used to train the residual feature set and construct the residual fitting model; The structure of the residual fitting model includes an input layer, a hidden layer and an output layer. The input layer receives the residual feature set, and the output layer outputs the corrected spraying position and thickness control results, and reduces the residual through the nonlinear fitting ability of the neural network.

7. The precise positioning and spraying thickness control system in an aircraft painting hangar according to claim 1, characterized in that: The data acquisition unit preprocesses the collected data in a manner including data denoising and normalization processing.

8. The precise positioning and spraying thickness control system in an aircraft painting hangar according to claim 1, characterized in that: Methods for denoising data include: A filtering algorithm is used to remove high-frequency noise from the data collected by the positioning sensor and the thickness measurement sensor; For abnormal data points caused by equipment failure, the average value of adjacent data points is used for replacement.

9. The precise positioning and spraying thickness control system in an aircraft painting hangar according to claim 1, characterized in that: The method of normalizing the data includes: performing linear transformation on the positioning data and the thickness data according to their maximum and minimum values ​​so that their value ranges are mapped to the interval [0,1].

Images