Pressure Control System and Method Based on Adaptive Airbag Regulation

Abstract

This invention relates to the technical field of pressure control, and discloses a pressure control system and method based on adaptive airbag adjustment. First, an initial airbag pressure data set is acquired, and this data set is cleaned and repaired to obtain a processed airbag pressure data set. Second, a backpropagation (BP) neural network is trained, and the parameters in the BP neural network are optimized using a Hunger Games search algorithm to obtain optimized parameters, thus establishing a BP neural network prediction model. Then, based on the processed airbag pressure data set and the BP neural network prediction model, predicted airbag pressure values ​​are output. These predicted values ​​are then divided according to pressure detection points to generate an airbag pressure data matrix. Finally, a threshold is set to achieve real-time early warning and regulation of airbag pressure, thereby preventing pressure ulcers. This invention achieves the purpose of airbag pressure control through processing and analysis of airbag pressure data, and the method is objective and accurate.

Description

Technical Field

[0001] This invention relates to the technical field of pressure control, specifically to a pressure control system and method based on adaptive airbag adjustment. Background Technology

[0002] Traditional methods of regulating airbag pressure typically rely on manual adjustment, making it difficult to precisely control the airbag pressure. This prevents the airbag from distributing pressure points and relieving localized pressure, thus failing to provide the massage effect that promotes blood circulation. Furthermore, the lack of artificial intelligence technology means that the airbag pressure adjustment is slow and cannot intelligently detect pressure points or automatically and selectively inflate surrounding airbags to relieve pressure, thus failing to prevent pressure sores and causing inconvenience for users. Summary of the Invention

[0003] In view of the problems in the related technologies, the present invention provides a pressure control system and method based on adaptive airbag adjustment to overcome the above-mentioned technical problems existing in the prior art.

[0004] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0005] This invention relates to a pressure control method based on adaptive airbag adjustment, comprising the following steps:

[0006] S1. Set pressure detection points on the airbag, set a time series to obtain an initial airbag pressure data set, perform data cleaning and data repair on the initial airbag pressure data set to obtain a processed airbag pressure data set.

[0007] S2. Reacquire the historical airbag pressure data set, train the BP neural network, and use the Hunger Games search algorithm to optimize the parameters in the BP neural network to obtain the optimized parameters. Use the optimized parameters to build a BP neural network prediction model.

[0008] S3. Based on the processed airbag pressure data set and the BP neural network prediction model, output the predicted airbag pressure data value, then obtain the predicted airbag pressure data value of the pressure detection point, and generate the airbag pressure data matrix.

[0009] S4. Set the pressure threshold and time threshold for the pressure detection point to achieve real-time early warning of airbag pressure, and then adjust the airbag pressure in real time according to the airbag pressure data matrix to prevent pressure sores.

[0010] This invention obtains an initial airbag pressure data set by setting pressure detection points on the airbag, and then performs data cleaning and repair to obtain a processed airbag pressure data set. This method improves the reliability of airbag pressure data by deleting abnormal data and repairs missing data based on data distance, greatly increasing the accuracy of subsequent data prediction and reducing pressure control errors. Secondly, a backpropagation (BP) neural network is trained, and the parameters in the BP neural network are optimized using the Hunger Games search algorithm to construct a BP neural network prediction model. This algorithm finds the optimal solution to the problem by simulating animal hunger and foraging behavior, exhibiting good solution accuracy and fast convergence performance. Its simple structure is suitable for model optimization, improving the prediction speed and accuracy of the prediction model, reducing the time consumed in airbag pressure control, and using the predicted values ​​for pressure analysis facilitates subsequent pressure warnings. Furthermore, based on the BP neural network prediction model, predicted airbag pressure data values ​​are output and divided to generate an airbag pressure data matrix. By assigning the predicted values ​​to corresponding pressure detection points, local pressure data analysis is facilitated, improving the feasibility and adaptability of airbag adjustment. Finally, pressure and time thresholds are set to achieve real-time warning and control of airbag pressure.

[0011] Preferably, step S1 includes the following steps:

[0012] S11. Obtain an airbag, the airbag being used to prevent pressure sores, and establish a Cartesian coordinate system on the airbag, performing a gridding process on the airbag to divide it into several sections of length [missing information]. Width is The airbag is constructed using a grid pattern. Pressure detection points are set inside the airbag grid, and pressure sensors are used to collect airbag pressure data at these detection points, generating a time series based on the collected data. ,in This represents the time point at which the nth airbag pressure data was collected; the pressure data are sorted according to the time series to generate the initial airbag pressure data set. ,in This represents the subset of airbag pressure data at the m-th pressure detection point;

[0013] S12. Select any subset of airbag pressure data from the initial airbag pressure data set, denoted as the airbag pressure data set to be processed. Set an upper pressure threshold and a lower pressure threshold. Delete the airbag pressure data in the airbag pressure data set to be processed that are greater than the upper pressure threshold and less than the lower pressure threshold, completing data cleaning and obtaining the processed airbag pressure data set. Then, perform data repair on the processed airbag pressure data set based on the Euclidean distance to obtain the properly processed airbag pressure data set. The specific steps are as follows:

[0014] S121. Select the first [time series] from the time series. Time points for collecting airbag pressure data and the +1 airbag pressure data acquisition time point Calculate the data collection time point and collection time point The correlations of the corresponding airbag pressure data are analyzed, and the correlations are sorted in descending order to obtain a correlation set. The airbag pressure data corresponding to the top k correlations in the correlation set are selected to form the first airbag pressure data sample set and the second airbag pressure data sample set. The Euclidean distance of the data is calculated based on the correlations and the airbag pressure data sample sets, and the calculation formula is as follows:

[0015] ;

[0016] Where c represents the Euclidean distance of the data. Indicates the correlation of airbag pressure data. This represents the i-th airbag pressure data in the first airbag pressure data sample set. This represents the i-th airbag pressure data in the second airbag pressure data sample set. ;

[0017] S122. Find the missing data in the processed airbag pressure data set, calculate the average value of the airbag pressure data within the Euclidean distance range of the missing data, and use the average value of the airbag pressure data to repair the missing data in the processed airbag pressure data set to obtain the processed airbag pressure data set.

[0018] This invention obtains an initial airbag pressure data set by setting pressure detection points on the airbag, and then performs data cleaning and repair. Data cleaning improves the reliability of airbag pressure data, and missing data is repaired according to data distance, which greatly increases the accuracy of subsequent data prediction, reduces pressure control errors, and obtains a processed airbag pressure data set.

[0019] Preferably, step S2 includes the following steps:

[0020] S21. Reacquire the historical airbag pressure data set. ,in This represents a subset of historical airbag pressure data for the m-th pressure detection point, and records the acquisition time of the historical airbag pressure data to generate a historical time series. ,in This represents the time point when the nth historical airbag pressure data was collected. The historical airbag pressure data set is then cleaned and repaired to obtain a historical airbag pressure data sample set.

[0021] S22. Construct a BP neural network, which includes an input layer, hidden layers, an output layer, and a target layer. Input the historical airbag pressure data sample set into the BP neural network. The input layer calculates weights and thresholds and passes them to the hidden layers. The hidden layers are mapped to the output layer through activation functions. The output layer is mapped to the target layer through activation functions. Calculate the target layer error and backpropagate to adjust the weights and thresholds. Treat the weights and thresholds as parameters in the BP neural network. Use the Hunger Games search algorithm to optimize the parameters in the BP neural network to obtain optimized parameters. The specific steps are as follows:

[0022] S221. Using the target layer error as the fitness function, construct a search space. An animal population of size r exists in the search space. Animals in the population are considered parameters in the BP neural network. When the animal population enters the food-approaching stage, initialize the animal population and obtain its initial position. The animal population forages at its initial position. Set the current iteration number to d and the convergence factor to... , Indicates that it lies within the interval [- , The random number. This represents a random number between the interval [0, 1]. and Indicates the hunger weight. Let represent a random number that follows a normal distribution. The optimal animal position in the population at the d-th iteration is denoted as . For the animal position at the d-th iteration Update; when At the time of the (d+1)th iteration, the animal's position ,when At the time of the (d+1)th iteration, the animal's position ;

[0023] S222. Calculate the worst fitness function value and the best fitness function value in the current iteration, denoted as follows: and The fitness function value of the g-th individual animal in the animal population is The upper and lower bounds of the search space are respectively and , This represents a random number between the interval [0, 1], where the foraging coefficient is... Set the hunger level of the g-th animal in the population to be... , This represents a random number between [0, 1]. When the foraging coefficient is less than 100, the hunger coefficient is... When the foraging coefficient is greater than or equal to 100, the hunger coefficient is... The hunger coefficient is used to update the hunger level to obtain a new hunger level. The hunger weight of the animal population entering the near-food stage is updated using the new hunger level to generate the next generation of animal population. The animal positions in the next generation of animal population are updated. A maximum number of iterations is set. The iteration stops when the current iteration count reaches the maximum number of iterations to obtain the final animal population. The optimal fitness function value is calculated, and the final animal position corresponding to the optimal fitness function value is found. The final animal position is regarded as the optimized parameter.

[0024] S23. The optimized parameters are used as the optimal weights and thresholds of the BP neural network. The BP neural network is continuously trained and iterated until it converges, thus establishing a BP neural network prediction model.

[0025] This invention trains a BP neural network and optimizes its parameters using the Hunger Games search algorithm. By simulating the hunger and foraging behavior of animals, it finds the optimal solution to the problem. It has good solution accuracy and fast convergence performance. Its simple structure is suitable for model optimization, which improves the prediction speed and accuracy of the prediction model. The constructed BP neural network prediction model reduces the time required for airbag pressure control and facilitates subsequent pressure warning.

[0026] Preferably, step S3 includes the following steps:

[0027] S31. Select a subset of airbag pressure data from the pressure detection points in the processed airbag pressure data set, and divide the airbag pressure data subset according to the time series corresponding to the pressure detection points, obtaining several airbag pressure data segments. These segments are then sequentially input into the BP neural network prediction model to output a set of predicted airbag pressure data values. ,in This represents the predicted airbag pressure data for the j-th airbag pressure data segment;

[0028] S32. Number the pressure detection points to obtain a set of pressure detection points. ,in Indicates the first There are several pressure detection points. Based on the set of pressure detection points, the predicted airbag pressure data set is classified, and the predicted airbag pressure data values ​​in the predicted airbag pressure data set are assigned to the corresponding pressure detection points to generate an airbag pressure data matrix, as follows:

[0029] ;

[0030] Where D represents the airbag pressure data matrix, Indicates the first Predicted airbag pressure data for the qth airbag at each pressure detection point.

[0031] This invention outputs predicted airbag pressure data based on a BP neural network prediction model. By dividing the predicted values ​​into corresponding pressure detection points, an airbag pressure data matrix is ​​generated, which facilitates the analysis of local pressure data and improves the feasibility and adaptability of airbag adjustment.

[0032] Preferably, step S4 includes the following steps:

[0033] S41. Based on the predicted airbag pressure data of the pressure detection points in the airbag pressure data matrix, set the pressure threshold and time threshold of the pressure detection points. Calculate the time when the predicted airbag pressure data is less than the pressure threshold of the pressure detection point according to the time series. Within the time period t, when the time when the predicted airbag pressure data is less than the pressure threshold of the pressure detection point is greater than the time threshold of the pressure detection point, the corresponding pressure detection point is recorded as an abnormal pressure detection point; otherwise, it is not an abnormal pressure detection point. An alarm is triggered at the abnormal pressure detection point to achieve real-time early warning of airbag pressure.

[0034] S42. After issuing an early warning at the abnormal pressure detection point, the predicted value of the airbag pressure data at the abnormal pressure detection point in the airbag pressure data matrix is ​​adjusted. The airbags around the abnormal pressure detection point are inflated to reduce the pressure at the abnormal pressure detection point, so that the abnormal pressure detection point is suspended and no longer under pressure, until the predicted value of the airbag pressure data at the abnormal pressure detection point is less than the pressure threshold of the pressure detection point, thereby preventing pressure sores and completing the real-time control of the airbag pressure.

[0035] This invention also discloses a system based on an adaptive airbag adjustment pressure control method, specifically including: a pressure data cleaning and repair module, a neural network prediction model construction module, a pressure data prediction value generation module, and an airbag pressure early warning and control module;

[0036] The pressure data cleaning and repair module is used to clean and repair the initial airbag pressure data set.

[0037] The neural network prediction model building module is used to optimize the parameters in the BP neural network using the Hunger Games search algorithm to establish a BP neural network prediction model.

[0038] The pressure data prediction generation module is used to output airbag pressure data prediction values ​​based on the model and to allocate the data.

[0039] The airbag pressure warning and control module is used to set thresholds to achieve real-time warning and control of airbag pressure.

[0040] The present invention has the following beneficial effects:

[0041] 1. This invention obtains an initial airbag pressure data set by setting pressure detection points on the airbag, and performs data cleaning and repair. Data cleaning improves the reliability of airbag pressure data, and missing data is repaired according to data distance, which greatly increases the accuracy of subsequent data prediction, reduces pressure control errors, and obtains a processed airbag pressure data set.

[0042] 2. This invention trains a BP neural network and optimizes the parameters in the BP neural network using the Hunger Games search algorithm. By simulating the hunger activities and foraging behaviors of animals, it finds the optimal solution to the problem. It has good solution accuracy and fast convergence performance. The simple structure is suitable for model optimization, which improves the prediction speed and prediction accuracy of the prediction model. The constructed BP neural network prediction model reduces the time consumed by airbag pressure control and facilitates subsequent pressure warning.

[0043] 3. This invention outputs predicted airbag pressure data based on a BP neural network prediction model. By dividing the predicted values ​​into corresponding pressure detection points, an airbag pressure data matrix is ​​generated, which facilitates the analysis of local pressure data and improves the feasibility and adaptability of airbag adjustment.

[0044] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of the pressure control process of the pressure control system based on adaptive airbag adjustment provided by the present invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] In the description of this invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inner", etc., which indicate orientation or positional relationship, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the invention.

[0049] Example 1

[0050] Please refer to Figure 1 This invention is a pressure control method based on adaptive airbag adjustment. Taking a segmented inflatable airbag as an example, it includes the following steps:

[0051] (1) Set up an inflatable air cushion (not a one-piece whole), inflate and raise it in a wave-like manner from far to far to promote venous return in the distal limbs, prevent deep vein thrombosis in the lower limbs, disperse pressure points on the body in a wave-like manner, relieve local pressure, and at the same time massage blood circulation to prevent pressure sores, combined with intelligent detection.

[0052] (2) Set local pressure threshold and local pressure duration threshold; when the local pressure of the inflatable air cushion is greater than or equal to the local pressure threshold and the duration is greater than or equal to the local pressure duration threshold, an alarm is triggered on the pressure point, and the surrounding airbags are automatically and selectively inflated to suspend the pressure area, relieve the pressure, and prevent pressure sores.

[0053] The above-mentioned alarm for pressure points includes the following steps:

[0054] S1. Set pressure detection points on the airbag, set a time series to obtain an initial airbag pressure data set, perform data cleaning and data repair on the initial airbag pressure data set to obtain a processed airbag pressure data set.

[0055] S1 includes the following steps:

[0056] S11. Obtain an airbag, the airbag being used to prevent pressure sores, and establish a Cartesian coordinate system on the airbag, performing a gridding process on the airbag to divide it into several sections of length [missing information]. Width is The airbag is constructed using a grid pattern. Pressure detection points are set inside the airbag grid, and pressure sensors are used to collect airbag pressure data at these detection points, generating a time series based on the collected data. ,in This represents the time point at which the nth airbag pressure data was collected; the pressure data are sorted according to the time series to generate the initial airbag pressure data set. ,in This represents the subset of airbag pressure data at the m-th pressure detection point;

[0057] S12. Select any subset of airbag pressure data from the initial airbag pressure data set, denoted as the airbag pressure data set to be processed. Set an upper pressure threshold and a lower pressure threshold. Delete the airbag pressure data in the airbag pressure data set to be processed that are greater than the upper pressure threshold and less than the lower pressure threshold, completing data cleaning and obtaining the processed airbag pressure data set. Then, perform data repair on the processed airbag pressure data set based on the Euclidean distance to obtain the properly processed airbag pressure data set. The specific steps are as follows:

[0058] S121. Select the first [time series] from the time series. Time points for collecting airbag pressure data and the +1 airbag pressure data acquisition time point Calculate the data collection time point and collection time point The correlations of the corresponding airbag pressure data are analyzed, and the correlations are sorted in descending order to obtain a correlation set. The airbag pressure data corresponding to the top k correlations in the correlation set are selected to form the first airbag pressure data sample set and the second airbag pressure data sample set. The Euclidean distance of the data is calculated based on the correlations and the airbag pressure data sample sets, and the calculation formula is as follows:

[0059] ;

[0060] Where c represents the Euclidean distance of the data. This indicates the correlation of airbag pressure data. This represents the i-th airbag pressure data in the first airbag pressure data sample set. This represents the i-th airbag pressure data in the second airbag pressure data sample set. ;

[0061] S122. Find the missing data in the processed airbag pressure data set, calculate the average value of the airbag pressure data within the Euclidean distance range of the missing data, and use the average value of the airbag pressure data to repair the missing data in the processed airbag pressure data set to obtain the processed airbag pressure data set.

[0062] S2. Reacquire the historical airbag pressure data set, train the BP neural network, and use the Hunger Games search algorithm to optimize the parameters in the BP neural network to obtain the optimized parameters. Use the optimized parameters to build a BP neural network prediction model.

[0063] S2 includes the following steps:

[0064] S21. Reacquire the historical airbag pressure data set. ,in This represents a subset of historical airbag pressure data for the m-th pressure detection point, and records the acquisition time of the historical airbag pressure data to generate a historical time series. ,in This represents the time point when the nth historical airbag pressure data was collected. The historical airbag pressure data set is then cleaned and repaired to obtain a historical airbag pressure data sample set.

[0065] S22. Construct a BP neural network, which includes an input layer, hidden layers, an output layer, and a target layer. Input the historical airbag pressure data sample set into the BP neural network. The input layer calculates weights and thresholds and passes them to the hidden layers. The hidden layers are mapped to the output layer through activation functions. The output layer is mapped to the target layer through activation functions. Calculate the target layer error and backpropagate to adjust the weights and thresholds. Treat the weights and thresholds as parameters in the BP neural network. Use the Hunger Games search algorithm to optimize the parameters in the BP neural network to obtain optimized parameters. The specific steps are as follows:

[0066] S221. Using the target layer error as the fitness function, construct a search space. An animal population of size r exists in the search space. Animals in the population are considered parameters in the BP neural network. When the animal population enters the food-approaching stage, initialize the animal population and obtain its initial position. The animal population forages at its initial position. Set the current iteration number to d and the convergence factor to... , Indicates that it lies within the interval [- , The random number. This represents a random number between the interval [0, 1]. and Indicates the hunger weight. Let represent a random number that follows a normal distribution. The optimal animal position in the population at the d-th iteration is denoted as . For the animal position at the d-th iteration Update; when At the time of the (d+1)th iteration, the animal's position ,when At the time of the (d+1)th iteration, the animal's position ;

[0067] S222. Calculate the worst fitness function value and the best fitness function value in the current iteration, denoted as follows: and The fitness function value of the g-th individual animal in the animal population is The upper and lower bounds of the search space are respectively and , This represents a random number between the interval [0, 1], where the foraging coefficient is... Set the hunger level of the g-th animal in the population to be... , This represents a random number between [0, 1]. When the foraging coefficient is less than 100, the hunger coefficient is... When the foraging coefficient is greater than or equal to 100, the hunger coefficient is... The hunger coefficient is used to update the hunger level to obtain a new hunger level. The hunger weight of the animal population entering the near-food stage is updated using the new hunger level to generate the next generation of animal population. The animal positions in the next generation of animal population are updated. A maximum number of iterations is set. The iteration stops when the current iteration count reaches the maximum number of iterations to obtain the final animal population. The optimal fitness function value is calculated, and the final animal position corresponding to the optimal fitness function value is found. The final animal position is regarded as the optimized parameter.

[0068] S23. The optimized parameters are used as the optimal weights and thresholds of the BP neural network. The BP neural network is continuously trained and iterated until the BP neural network converges, thus establishing a BP neural network prediction model.

[0069] S3. Based on the processed airbag pressure data set and the BP neural network prediction model, output the predicted airbag pressure data value, then obtain the predicted airbag pressure data value of the pressure detection point, and generate the airbag pressure data matrix.

[0070] S3 includes the following steps:

[0071] S31. Select a subset of airbag pressure data from the pressure detection points in the processed airbag pressure data set, and divide the airbag pressure data subset according to the time series corresponding to the pressure detection points, obtaining several airbag pressure data segments. These segments are then sequentially input into the BP neural network prediction model to output a set of predicted airbag pressure data values. ,in This represents the predicted airbag pressure data for the j-th airbag pressure data segment;

[0072] S32. Number the pressure detection points to obtain a set of pressure detection points. ,in Indicates the first There are several pressure detection points. Based on the set of pressure detection points, the predicted airbag pressure data set is classified, and the predicted airbag pressure data values ​​in the predicted airbag pressure data set are assigned to the corresponding pressure detection points to generate an airbag pressure data matrix, as follows:

[0073] ;

[0074] Where D represents the airbag pressure data matrix, Indicates the first Predicted airbag pressure data at pressure detection points q;

[0075] S4. Set the pressure threshold and time threshold for the pressure detection point to realize real-time early warning of airbag pressure, and then adjust the airbag pressure in real time according to the airbag pressure data matrix to prevent pressure sores.

[0076] S4 includes the following steps:

[0077] S41. Based on the predicted airbag pressure data of the pressure detection points in the airbag pressure data matrix, set the pressure threshold and time threshold of the pressure detection points. Calculate the time when the predicted airbag pressure data is less than the pressure threshold of the pressure detection point according to the time series. Within the time period t, when the time when the predicted airbag pressure data is less than the pressure threshold of the pressure detection point is greater than the time threshold of the pressure detection point, the corresponding pressure detection point is recorded as an abnormal pressure detection point; otherwise, it is not an abnormal pressure detection point. An alarm is triggered at the abnormal pressure detection point to achieve real-time early warning of airbag pressure.

[0078] S42. After issuing an early warning at the abnormal pressure detection point, the predicted value of the airbag pressure data at the abnormal pressure detection point in the airbag pressure data matrix is ​​adjusted. The airbags around the abnormal pressure detection point are inflated to reduce the pressure at the abnormal pressure detection point, so that the abnormal pressure detection point is suspended and no longer under pressure, until the predicted value of the airbag pressure data at the abnormal pressure detection point is less than the pressure threshold of the pressure detection point, thereby preventing pressure sores and completing the real-time control of the airbag pressure.

[0079] This invention also discloses a system based on an adaptive airbag adjustment pressure control method, specifically including: a pressure data cleaning and repair module, a neural network prediction model construction module, a pressure data prediction value generation module, and an airbag pressure early warning and control module;

[0080] The pressure data cleaning and repair module is used to clean and repair the initial airbag pressure data set.

[0081] The neural network prediction model building module is used to optimize the parameters in the BP neural network using the Hunger Games search algorithm to establish a BP neural network prediction model.

[0082] The pressure data prediction generation module is used to output airbag pressure data prediction values ​​based on the model and to allocate the data.

[0083] The airbag pressure warning and control module is used to set thresholds to achieve real-time warning and control of airbag pressure.

[0084] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0085] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A pressure control method based on adaptive airbag adjustment, characterized in that, Includes the following steps: S1. Set pressure detection points on the airbag, set a time series to obtain an initial airbag pressure data set, perform data cleaning and data repair on the initial airbag pressure data set to obtain a processed airbag pressure data set. S2. Reacquire the historical airbag pressure data set, train the BP neural network, and use the Hunger Games search algorithm to optimize the parameters in the BP neural network to obtain the optimized parameters. Use the optimized parameters to build a BP neural network prediction model. S3. Based on the processed airbag pressure data set and the BP neural network prediction model, output the predicted airbag pressure data value, then obtain the predicted airbag pressure data value of the pressure detection point, and generate the airbag pressure data matrix. Specifically, this includes: selecting a subset of airbag pressure data from pressure detection points in the processed airbag pressure data set, inputting it sequentially into a BP neural network prediction model, and outputting a set of predicted airbag pressure data values; dividing the predicted airbag pressure data values ​​in the set of predicted airbag pressure data values ​​into the corresponding pressure detection points, and generating an airbag pressure data matrix. S4. Set the pressure threshold and time threshold for the pressure detection point to realize real-time early warning of airbag pressure, and then adjust the airbag pressure in real time according to the airbag pressure data matrix to prevent pressure sores. Specifically, this includes: setting pressure thresholds and time thresholds for pressure detection points based on the predicted airbag pressure data values ​​at pressure detection points in the airbag pressure data matrix; calculating the time when the predicted airbag pressure data value is less than the pressure threshold for the pressure detection point based on the time series; within the time period t, when the time when the predicted airbag pressure data value is less than the pressure threshold for the pressure detection point is greater than the time threshold for the pressure detection point, the corresponding pressure detection point is recorded as an abnormal pressure detection point; otherwise, it is not an abnormal pressure detection point, and an alarm is triggered at the abnormal pressure detection point to achieve real-time early warning of airbag pressure.

2. The pressure control method based on adaptive airbag adjustment according to claim 1, characterized in that, S1 includes the following steps: S11. Obtain the airbag and perform grid processing on the airbag to obtain the airbag grid. Set pressure detection points inside the airbag grid, and then use pressure sensors to collect the airbag pressure data at the pressure detection points. Obtain the collection time points to form a time series and obtain the initial airbag pressure data set. S12. Select any subset of airbag pressure data from the initial airbag pressure data set, and denote it as the airbag pressure data set to be processed; set an upper pressure threshold and a lower pressure threshold, and compare the airbag pressure data set to be processed with the upper and lower pressure thresholds to achieve data cleaning and obtain the processed airbag pressure data set; then perform data repair on the processed airbag pressure data set based on the Euclidean distance of the data to obtain the processed airbag pressure data set.

3. The pressure control method based on adaptive airbag adjustment according to claim 2, characterized in that, S12 includes the following steps: S121. Select the data collection time points in the time series, calculate the correlation of the airbag pressure data corresponding to the data collection time points, obtain the correlation set, and select the first data points in the correlation set. k The airbag pressure data corresponding to each correlation are used to form an airbag pressure data sample set; the Euclidean distance of the data is calculated based on the correlation and the airbag pressure data sample set. S122. Find the missing data in the processed airbag pressure data set, calculate the average value of the airbag pressure data within the Euclidean distance range of the missing data, and use the average value of the airbag pressure data to repair the missing data in the processed airbag pressure data set to obtain the processed airbag pressure data set.

4. The pressure control method based on adaptive airbag adjustment according to claim 3, characterized in that, S2 includes the following steps: S21. Reacquire the historical airbag pressure data set, and perform data cleaning and data repair on the historical airbag pressure data set to obtain a historical airbag pressure data sample set. S22. Construct a BP neural network, which includes an input layer, hidden layers, an output layer, and a target layer. Input the historical airbag pressure data sample set into the BP neural network. The input layer calculates weights and thresholds and passes them to the hidden layers. The hidden layers are mapped to the output layer through activation functions. The output layer is mapped to the target layer through activation functions. Calculate the target layer error and backpropagate to adjust the weights and thresholds. Treat the weights and thresholds as parameters in the BP neural network, and use the Hunger Games search algorithm to optimize the parameters in the BP neural network to obtain optimized parameters. S23. The optimized parameters are used as the optimal weights and thresholds of the BP neural network. The BP neural network is continuously trained and iterated until it converges, thus establishing a BP neural network prediction model.

5. The pressure control method based on adaptive airbag adjustment according to claim 4, characterized in that, S22 includes the following steps: S221. Using the target layer error as the fitness function, construct a search space. There is an animal population in the search space, and the number of animal populations is r. The animals in the animal population are regarded as parameters in the BP neural network. The animal population enters the near-food stage, initializes the population, obtains its initial position, and forages at the initial position. The current iteration number is set to [number missing]. d The convergence factor is , Indicates that it lies within the interval [- , The random number. This represents a random number within the interval [0, 1]. and Indicates the hunger weight. Represents a random number that follows a normal distribution, selected as the th... d The optimal animal position in the population at the next iteration is denoted as... , for the d Animal position at the next iteration Update; when At that time, the first d Animal position at +1 iteration ,when At that time, the first d Animal position at +1 iteration ; S222. Calculate the worst fitness function value and the best fitness function value in the current iteration, denoted as follows: and The first in the animal population g The fitness function value of an individual animal is The upper and lower bounds of the search space are respectively and , This represents a random number between the interval [0, 1], where the foraging coefficient is... ; Set the first in the animal population g The hunger level of each individual animal is , This represents a random number between [0, 1]. When the foraging coefficient is less than 100, the hunger coefficient is... When the foraging coefficient is greater than or equal to 100, the hunger coefficient is... The hunger coefficient is used to update the hunger level to obtain a new hunger level. The hunger weight of the animal population entering the near-food stage is updated using the new hunger level to generate the next generation of animal population. The animal positions in the next generation of animal population are updated. A maximum number of iterations is set. The iteration stops when the current iteration count reaches the maximum number of iterations to obtain the final animal population. The optimal fitness function value is calculated, and the final animal position corresponding to the optimal fitness function value is found. The final animal position is regarded as the optimized parameter.

6. The pressure control method based on adaptive airbag adjustment according to claim 5, characterized in that, S4 includes the following steps: S41. Set the pressure threshold and time threshold for the pressure detection point, and issue an early warning at the abnormal pressure detection point to realize real-time early warning of airbag pressure. S42. After issuing an early warning at the abnormal pressure detection point, the predicted value of the airbag pressure data at the abnormal pressure detection point in the airbag pressure data matrix is ​​adjusted. The airbags around the abnormal pressure detection point are inflated to reduce the pressure at the abnormal pressure detection point, so that the abnormal pressure detection point is suspended and no longer under pressure, until the predicted value of the airbag pressure data at the abnormal pressure detection point is less than the pressure threshold of the pressure detection point, thereby preventing pressure sores and completing the real-time control of the airbag pressure.

7. A system for implementing the pressure control method based on adaptive airbag adjustment as described in any one of claims 1-6, characterized in that, Specifically, it includes: The system includes a pressure data cleaning and repair module, a neural network prediction model construction module, a pressure data prediction value generation module, and an airbag pressure early warning and control module. The pressure data cleaning and repair module is used to clean and repair the initial airbag pressure data set. The neural network prediction model building module is used to optimize the parameters in the BP neural network using the Hunger Games search algorithm to establish a BP neural network prediction model. The pressure data prediction generation module is used to output airbag pressure data prediction values ​​based on the model and to allocate the data. The airbag pressure warning and control module is used to set thresholds to achieve real-time warning and control of airbag pressure.

Images