Shell state recognition method based on local optimization and support vector machine

Abstract

The invention relates to the technical field of shell state recognition, and discloses a shell state recognition method based on local optimization and a support vector machine, and the method comprises the steps: collecting state data of an underwater vehicle shell, the state data comprising stress data, deformation data and acoustic emission data; key features are extracted based on a local optimization algorithm; training a support vector machine model according to the extracted key features to obtain a risk value of the shell state; acquiring driving data of the underwater vehicle, adjusting the risk value according to the driving data, and determining a final risk value of the shell; and generating a driving strategy of the underwater vehicle according to the final risk value. By analyzing and processing the shell state data and combining the local optimization algorithm and the support vector machine model, the accuracy and efficiency of shell state recognition are improved, real-time monitoring of the shell state of the underwater vehicle is achieved, and potential safety problems can be found in time conveniently.

Description

Technical Field

[0001] The invention relates to the technical field of shell state recognition, and in particular to a shell state recognition method based on local optimization and support vector machine. Background Art

[0002] With the development of marine resources and the deepening of marine scientific research, underwater vehicles (such as submarines, unmanned submersibles, etc.) are important detection tools, and their safety performance and reliability are of vital importance. The shell of an underwater vehicle is a key structural part that protects internal instruments and equipment from seawater pressure and external environment. When performing tasks, it often faces extreme marine environments, which may cause damage or deterioration of its shell structure, thereby affecting its performance and safety. Therefore, monitoring the health status of the shell is of great significance to ensure the safe operation of underwater vehicles.

[0003] Traditional hull monitoring methods mainly rely on regular inspections and manual visual inspections, which are not only inefficient but also fail to achieve real-time monitoring and early warning of the hull status. In addition, due to the complexity of the underwater environment, manual inspections often have difficulty accurately assessing the damage degree and location of the hull. Therefore, there is an urgent need for a hull status recognition method based on local optimization and support vector machines to achieve real-time monitoring and recognition of the underwater vehicle hull. Summary of the invention

[0004] The purpose of the present invention is to provide a shell state recognition method based on local optimization and support vector machine, which solves the problem that traditional shell monitoring mainly relies on regular inspection and manual visual inspection, which is not only inefficient but also unable to achieve real-time monitoring and early warning of the shell state. In addition, due to the complexity of the underwater environment, manual inspection is often difficult to accurately assess the damage degree and location of the shell.

[0005] The present invention provides a shell state recognition method based on local optimization and support vector machine, the method comprising:

[0006] Collecting state data of the underwater vehicle shell, the state data including stress data, deformation data and acoustic emission data:

[0007] Processing the state data and extracting key features based on a local optimization algorithm;

[0008] The support vector machine model is trained based on the extracted key features to obtain the risk value of the shell state;

[0009] Acquiring driving data of the underwater vehicle, adjusting the risk value according to the driving data, and determining a final risk value of the shell;

[0010] A driving strategy for the underwater vehicle is generated according to the final risk value.

[0011] In some embodiments of the present application, the state data is processed and key features are extracted based on a local optimization algorithm, including:

[0012] Perform filtering, denoising and normalization on the collected data;

[0013] Extract features from the processed state data to obtain a feature set;

[0014] The feature set is optimized using a local optimization algorithm to find the best data combination and obtain key features.

[0015] In some embodiments of the present application, a support vector machine model is trained according to the extracted key features to obtain a risk value of the shell state, including:

[0016] The status data is historical status data and real-time status data;

[0017] Dividing the key features into a training set and a test set;

[0018] The support vector machine model is trained according to the training set, and the optimal parameters of the support vector machine model are found by a cross-validation method to establish the support vector machine model;

[0019] Verifying the support vector machine model according to the test set to determine the accuracy of the support vector machine model;

[0020] When the accuracy reaches a preset accuracy, a risk value F of the shell state is determined according to the real-time state data.

[0021] In some embodiments of the present application, the risk value is adjusted according to the driving data to obtain a final risk value of the housing, including:

[0022] The driving data includes driving speed and diving depth;

[0023] Comparing the driving speed with a speed threshold, if the driving speed is less than or equal to the speed threshold, not adjusting the risk value, and if the driving speed is greater than the speed threshold, adjusting the risk value according to a speed difference between the driving speed and the speed threshold;

[0024] The diving depth is compared with the depth threshold. If the diving depth is less than or equal to the depth threshold, the risk value is not adjusted. If the diving depth is greater than the depth threshold, the risk value is adjusted according to the depth difference between the diving depth and the depth threshold to obtain the final risk value of the shell.

[0025] In some embodiments of the present application, adjusting the risk value according to the speed difference between the driving speed and the speed threshold includes:

[0026] Determine a speed difference V0 between the driving speed V and the speed threshold Vm, V0 = V-Vm;

[0027] Presetting a first preset speed difference, a second preset speed difference and a third preset speed difference, wherein the first preset speed difference, the second preset speed difference and the third preset speed difference increase in sequence;

[0028] Correcting the risk value F according to the relationship between the speed difference V0 and the first preset speed difference, the second preset speed difference and the third preset speed difference;

[0029] If the speed difference V0 is less than the first preset speed difference, the first speed correction coefficient m1 is selected to correct the risk value F, and the corrected value is F×m1;

[0030] If the speed difference V0 is greater than or equal to the first preset speed difference, and the speed difference V0 is less than the second preset speed difference, then the second speed correction coefficient m2 is selected to correct the risk value F, and the corrected value is F×m2;

[0031] If the speed difference V0 is greater than or equal to the second preset speed difference, and the speed difference V0 is less than the third preset speed difference, then the third speed correction coefficient m3 is selected to correct the risk value F, and the corrected value is F×m3;

[0032] If the speed difference V0 is greater than or equal to the third preset speed difference, the fourth speed correction coefficient m4 is selected to correct the risk value F, and the corrected value is F×m4; wherein 1<m1<m2<m3<m4<1.2.

[0033] In some embodiments of the present application, adjusting the risk value according to the depth difference between the diving depth and the depth threshold to obtain a final risk value of the shell includes:

[0034] Determine the depth difference H0 between the diving depth H and the depth threshold Hm, H0 = H-Hm;

[0035] If the driving speed is less than or equal to the speed threshold, the risk value F is adjusted according to the depth difference H0 to obtain a final risk value of the shell;

[0036] If the driving speed is greater than the speed threshold, the corrected F×mi is adjusted according to the depth difference H0, i=1, 2, 3, 4, to obtain the final risk value of the shell.

[0037] In some embodiments of the present application, the risk value F is adjusted according to the depth difference H0 to obtain a final risk value of the shell, including:

[0038] Presetting a first preset depth difference, a second preset depth difference and a third preset depth difference, wherein the first preset depth difference, the second preset depth difference and the third preset depth difference increase in sequence;

[0039] adjusting the risk value F according to the relationship between the depth difference and the first preset depth difference, the second preset depth difference and the third preset depth difference;

[0040] If the depth difference H0 is less than the first preset depth difference, a first depth correction coefficient n1 is selected to correct the risk value F, so that a final risk value is obtained as F×n1;

[0041] If the depth difference H0 is greater than or equal to the first preset depth difference, and the depth difference H0 is less than the second preset depth difference, then a second depth correction coefficient n2 is selected to correct the risk value F, so that a final risk value is obtained as F×n2;

[0042] If the depth difference H0 is greater than or equal to the second preset depth difference, and the depth difference H0 is less than the third preset depth difference, a third depth correction coefficient n3 is selected to correct the risk value F, so as to obtain a final risk value of F×n3;

[0043] If the depth difference H0 is greater than or equal to the third preset depth difference, the fourth depth correction coefficient n4 is selected to correct the risk value F to obtain a final risk value of F×n4; wherein 1<n1<n2<n3<n4<1.2.

[0044] In some embodiments of the present application, the modified F×mi is adjusted according to the depth difference H0 to obtain the final risk value of the shell, including:

[0045] adjusting the corrected F×mi according to the relationship between the depth difference and the first preset depth difference, the second preset depth difference and the third preset depth difference;

[0046] If the depth difference H0 is less than the first preset depth difference, the fifth depth correction coefficient n5 is selected to correct the corrected F×mi to obtain a final risk value of F×mi×n5;

[0047] If the depth difference H0 is greater than or equal to the first preset depth difference, and the depth difference H0 is less than the second preset depth difference, a sixth depth correction coefficient n6 is selected to correct the risk value F, and a final risk value is obtained as F×mi×n6;

[0048] If the depth difference H0 is greater than or equal to the second preset depth difference, and the depth difference H0 is less than the third preset depth difference, then the seventh depth correction coefficient n7 is selected to correct the risk value F, and a final risk value is obtained as F×mi×n7;

[0049] If the depth difference H0 is greater than or equal to the third preset depth difference, the eighth depth correction coefficient n8 is selected to correct the risk value F to obtain a final risk value of F×mi×n8; wherein 1<n5<n6<n7<n8<1.2, and n1<n5, n2<n6, n3<n7, n4<n8.

[0050] In some embodiments of the present application, generating a driving strategy for the underwater vehicle according to the final risk value includes:

[0051] Presetting a first preset final risk value and a second preset final risk value, wherein the first preset final risk value is smaller than the second preset final risk value;

[0052] When the final risk value is less than the first preset final risk value, the driving strategy of the underwater vehicle is generated to be that the underwater vehicle can continue to drive;

[0053] When the final risk value is greater than or equal to the first preset final risk value, and the final risk value is less than the second preset final risk value, a driving strategy of the underwater vehicle is generated that the underwater vehicle can continue to drive, but the driving time does not exceed the preset time;

[0054] When the final risk value is greater than or equal to the second preset final risk value, the driving strategy of the underwater vehicle is generated to stop driving and return immediately.

[0055] Compared with the prior art, the beneficial effect of the present invention lies in that, by analyzing and processing the hull status data and combining the local optimization algorithm and the support vector machine model, the present invention improves the accuracy and efficiency of hull status identification, realizes real-time monitoring of the hull status of the underwater vehicle, facilitates timely discovery of potential safety problems, and adopts online monitoring methods to reduce dependence on manual inspection and reduce maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0057] Figure 1 It is a flow chart of the shell state recognition method based on local optimization and support vector machine of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0059] like Figure 1 As shown, the present invention provides a shell state recognition method based on local optimization and support vector machine, the method comprising:

[0060] Step S1, collecting status data of the underwater vehicle hull, wherein the status data includes stress data, deformation data and acoustic emission data.

[0061] It is understandable that a variety of sensors are set on the hull of the underwater vehicle to collect the status data of the hull of the underwater vehicle. Stress is the internal force per unit area. When the hull of the underwater vehicle is subjected to external forces (such as water pressure, collision, load, etc.), internal forces will be generated inside it to resist these external forces. The distribution of these internal forces inside the structure is called stress, which can be directly measured using stress sensors. Deformation refers to the change in shape or volume of a structure under the action of external forces, which is accompanied by stress. When the shell is subjected to external forces and generates stress, the material will undergo elastic deformation or plastic deformation. Elastic deformation can be restored to its original state after the external force is removed, while plastic deformation is permanent. Deformation can be measured by displacement sensors, strain gauges and other equipment, and the data usually includes displacement, strain size, etc. Acoustic emission refers to transient elastic waves generated when local changes occur inside the material (such as the formation and expansion of cracks, plastic deformation, etc.). When small structural changes occur inside the shell, energy is released rapidly. This energy propagates to the surface of the material in the form of waves and is captured by the acoustic emission sensor. Acoustic emission data is usually a series of electrical signals that are bursty in time, and each signal contains information about the location, intensity, and nature of the event. Acoustic emission data can be used to detect and locate potential microscopic damage.

[0062] Step S2, processing the state data and extracting key features based on a local optimization algorithm.

[0063] In some embodiments of the present application, the state data is processed and key features are extracted based on a local optimization algorithm, including: filtering, denoising and normalizing the collected data; extracting features from the processed state data to obtain a feature set; optimizing the feature set using a local optimization algorithm to find the best data combination to obtain key features.

[0064] It is understandable that the collected state data is raw data, and a series of processing is required to improve the data quality. Since there are electrical signals in the collected data, it is necessary to filter it. The purpose of filtering is to remove unnecessary frequency components in the signal. Common filtering methods include low-pass filtering, high-pass filtering, band-pass filtering, etc. In practical applications, a suitable filter can be selected for preprocessing according to the characteristics of the signal and the analysis requirements. For example, a moving average filter can be used to smooth data and reduce the impact of random noise. The purpose of denoising is to eliminate noise in the signal and retain useful information. Wavelet transform is a commonly used denoising method, which can decompose the signal into sub-band coefficients of different frequencies and then remove noise through threshold processing. For example, the shift-invariant GHM multiwavelet transform can effectively remove noise while retaining the characteristic information of the data. The specific steps include pre-filtering, multiwavelet decomposition, threshold processing, multiwavelet reconstruction and post-filtering. Normalization is to scale the data so that it falls into a small specific interval, usually between 0 and 1. This can eliminate the impact of the data dimension, make different features comparable, and also help to speed up the convergence of machine learning algorithms. Commonly used normalization methods include minimum-maximum normalization and Z-score normalization.

[0065] Step S3, training a support vector machine model according to the extracted key features to obtain a risk value of the shell state.

[0066] In some embodiments of the present application, a support vector machine model is trained based on the extracted key features to obtain a risk value of the shell state, including: the state data, historical state data, and real-time state data; the key features are divided into a training set and a test set; the support vector machine model is trained based on the training set, and the optimal parameters of the support vector machine model are found by a cross-validation method to establish a support vector machine model; the support vector machine model is verified based on the test set to determine the accuracy of the support vector machine model; when the accuracy reaches a preset accuracy, the risk value F of the shell state is determined based on the real-time state data.

[0067] It is understandable that the status data includes historical status data and real-time status data. The historical status data is used to build a support vector machine model, and the real-time status data is used to obtain the shell status risk value. The optimal parameters of the support vector machine model are obtained based on the cross-validation method to ensure the accuracy of identifying the risk value.

[0068] Step S4, obtaining the travel data of the underwater vehicle, adjusting the risk value according to the travel data, and determining the final risk value of the shell.

[0069] It is understandable that the driving data is the driving speed and diving depth of the underwater vehicle, which can be measured using relevant sensors.

[0070] In some embodiments of the present application, the risk value is adjusted according to the driving data to obtain a final risk value of the shell, including: the driving data includes a driving speed and a diving depth.

[0071] The driving speed is compared with a speed threshold. If the driving speed is less than or equal to the speed threshold, the risk value is not adjusted. If the driving speed is greater than the speed threshold, the risk value is adjusted according to a speed difference between the driving speed and the speed threshold.

[0072] The diving depth is compared with the depth threshold. If the diving depth is less than or equal to the depth threshold, the risk value is not adjusted. If the diving depth is greater than the depth threshold, the risk value is adjusted according to the depth difference between the diving depth and the depth threshold to obtain the final risk value of the shell.

[0073] In some embodiments of the present application, the risk value is adjusted according to the speed difference between the driving speed and the speed threshold, including: determining a speed difference V0 between the driving speed V and the speed threshold Vm, V0=V-Vm; presetting a first preset speed difference, a second preset speed difference and a third preset speed difference, the first preset speed difference, the second preset speed difference and the third preset speed difference increasing in sequence; correcting the risk value F according to the relationship between the speed difference V0 and the first preset speed difference, the second preset speed difference and the third preset speed difference; if the speed difference V0 is less than the first preset speed difference, selecting a first speed correction coefficient m1 to correct the risk value F, and correcting After correction, it is F×m1; if the speed difference V0 is greater than or equal to the first preset speed difference, and the speed difference V0 is less than the second preset speed difference, the second speed correction coefficient m2 is selected to correct the risk value F, and the correction is F×m2; if the speed difference V0 is greater than or equal to the second preset speed difference, and the speed difference V0 is less than the third preset speed difference, the third speed correction coefficient m3 is selected to correct the risk value F, and the correction is F×m3; if the speed difference V0 is greater than or equal to the third preset speed difference, the fourth speed correction coefficient m4 is selected to correct the risk value F, and the correction is F×m4; wherein, 1<m1<m2<m3<m4<1.2.

[0074] In this embodiment, the hull state is related to the travel speed of the underwater vehicle. When the travel speed is higher, the pressure or stress on the underwater vehicle is also greater. Therefore, the risk value of the hull state of the underwater vehicle is adjusted according to the travel speed. When the travel speed is higher, the selected speed correction coefficient is also larger, and the corrected value is larger.

[0075] In some embodiments of the present application, the risk value is adjusted according to the depth difference between the diving depth and the depth threshold to obtain the final risk value of the shell, including: determining the depth difference H0 between the diving depth H and the depth threshold Hm, H0=H-Hm; if the driving speed is less than or equal to the speed threshold, adjusting the risk value F according to the depth difference H0 to obtain the final risk value of the shell; if the driving speed is greater than the speed threshold, adjusting the corrected F×mi according to the depth difference H0, i=1, 2, 3, 4, to obtain the final risk value of the shell.

[0076] In this embodiment, when the risk value is adjusted according to the diving depth, it is necessary to determine whether the risk value has been adjusted once in the previous step. When the driving speed is less than or equal to the speed threshold, the risk value has not been adjusted once. When the driving speed is greater than the speed threshold, it means that the risk value has been adjusted once. When the risk value is adjusted according to the diving depth again, the adjustment is made based on the corrected F×mi.

[0077] In some embodiments of the present application, the risk value F is adjusted according to the depth difference H0 to obtain a final risk value of the shell, including: presetting a first preset depth difference, a second preset depth difference and a third preset depth difference, wherein the first preset depth difference, the second preset depth difference and the third preset depth difference increase in sequence; adjusting the risk value F according to the relationship between the depth difference and the first preset depth difference, the second preset depth difference and the third preset depth difference; if the depth difference H0 is less than the first preset depth difference, selecting a first depth correction coefficient n1 to correct the risk value F to obtain a final risk value of F×n1; if the depth difference H0 is greater than or equal to At the first preset depth difference, and the depth difference H0 is less than the second preset depth difference, the second depth correction coefficient n2 is selected to correct the risk value F, and the final risk value is F×n2; if the depth difference H0 is greater than or equal to the second preset depth difference, and the depth difference H0 is less than the third preset depth difference, the third depth correction coefficient n3 is selected to correct the risk value F, and the final risk value is F×n3; if the depth difference H0 is greater than or equal to the third preset depth difference, the fourth depth correction coefficient n4 is selected to correct the risk value F, and the final risk value is F×n4; wherein, 1<n1<n2<n3<n4<1.2.

[0078] In some embodiments of the present application, the corrected F×mi is adjusted according to the depth difference H0 to obtain a final risk value of the shell, including: adjusting the corrected F×mi according to the relationship between the depth difference and the first preset depth difference, the second preset depth difference and the third preset depth difference; if the depth difference H0 is less than the first preset depth difference, selecting the fifth depth correction coefficient n5 to correct the corrected F×mi to obtain a final risk value of F×mi×n5; if the depth difference H0 is greater than or equal to the first preset depth difference, and the depth difference H0 is less than the second preset depth difference, selecting the sixth depth correction coefficient n6 corrects the risk value F to obtain a final risk value of F×mi×n6; if the depth difference H0 is greater than or equal to the second preset depth difference, and the depth difference H0 is less than the third preset depth difference, the seventh depth correction coefficient n7 is selected to correct the risk value F to obtain a final risk value of F×mi×n7; if the depth difference H0 is greater than or equal to the third preset depth difference, the eighth depth correction coefficient n8 is selected to correct the risk value F to obtain a final risk value of F×mi×n8; wherein, 1<n5<n6<n7<n8<1.2, and n1<n5, n2<n6, n3<n7, n4<n8.

[0079] Step S5: generating a driving strategy for the underwater vehicle according to the final risk value.

[0080] In some embodiments of the present application, a driving strategy for an underwater vehicle is generated based on the final risk value, including: presetting a first preset final risk value and a second preset final risk value, the first preset final risk value being less than the second preset final risk value; when the final risk value is less than the first preset final risk value, the driving strategy for the underwater vehicle is generated to be that it can continue driving; when the final risk value is greater than or equal to the first preset final risk value, and the final risk value is less than the second preset final risk value, the driving strategy for the underwater vehicle is generated to be that it can continue driving, but the driving time does not exceed the preset time; when the final risk value is greater than or equal to the second preset final risk value, the driving strategy for the underwater vehicle is generated to stop driving and return immediately.

[0081] In this embodiment, a driving strategy for the underwater vehicle is generated according to the final risk value. When the final risk value is too small, that is, the final risk value is less than the first preset final risk value, it means that the underwater vehicle can continue to drive; when the final risk value is greater than or equal to the first preset final risk value, and the final risk value is less than the second preset final risk value, it means that the state of the underwater vehicle shell has not reached an unbearable level and can continue to drive, but the driving time cannot be too long; when the final risk value is too large, that is, the final risk value is greater than or equal to the second preset final risk value, it means that the underwater vehicle is not suitable for continuing to drive and needs to return immediately for inspection.

[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

[0083] The system provided in the above embodiment is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be decomposed or combined. For example, the modules in the above embodiment can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not regarded as improper limitations of the present invention.

[0084] Those skilled in the art should be able to appreciate that the modules and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the technical field. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

Claims

1. A shell state recognition method based on local optimization and support vector machine, characterized in that: The method comprises: Collecting state data of the underwater vehicle shell, the state data including stress data, deformation data and acoustic emission data: Processing the state data and extracting key features based on a local optimization algorithm; The support vector machine model is trained based on the extracted key features to obtain the risk value of the shell state; Acquiring driving data of the underwater vehicle, adjusting the risk value according to the driving data, and determining a final risk value of the shell; A driving strategy for the underwater vehicle is generated according to the final risk value.

2. The shell state recognition method based on local optimization and support vector machine according to claim 1 is characterized in that: The state data is processed and key features are extracted based on a local optimization algorithm, including: Perform filtering, denoising and normalization on the collected data; Extract features from the processed state data to obtain a feature set; The feature set is optimized using a local optimization algorithm to find the best data combination and obtain key features.

3. The shell state recognition method based on local optimization and support vector machine according to claim 2 is characterized in that: The support vector machine model is trained based on the extracted key features to obtain the risk value of the shell status, including: The status data is historical status data and real-time status data; Dividing the key features into a training set and a test set; The support vector machine model is trained according to the training set, and the optimal parameters of the support vector machine model are found by a cross-validation method to establish the support vector machine model; Verifying the support vector machine model according to the test set to determine the accuracy of the support vector machine model; When the accuracy reaches a preset accuracy, a risk value F of the shell state is determined according to the real-time state data.

4. The shell state recognition method based on local optimization and support vector machine according to claim 3 is characterized in that: The risk value is adjusted according to the driving data to obtain a final risk value of the housing, including: The driving data includes driving speed and diving depth; Comparing the driving speed with a speed threshold, if the driving speed is less than or equal to the speed threshold, not adjusting the risk value, and if the driving speed is greater than the speed threshold, adjusting the risk value according to a speed difference between the driving speed and the speed threshold; The diving depth is compared with the depth threshold. If the diving depth is less than or equal to the depth threshold, the risk value is not adjusted. If the diving depth is greater than the depth threshold, the risk value is adjusted according to the depth difference between the diving depth and the depth threshold to obtain the final risk value of the shell.

5. The shell state recognition method based on local optimization and support vector machine according to claim 4 is characterized in that: The risk value is adjusted according to a speed difference between the driving speed and the speed threshold, comprising: Determine a speed difference V0 between the driving speed V and the speed threshold Vm, V0 = V-Vm; Presetting a first preset speed difference, a second preset speed difference and a third preset speed difference, wherein the first preset speed difference, the second preset speed difference and the third preset speed difference increase in sequence; Correcting the risk value F according to the relationship between the speed difference V0 and the first preset speed difference, the second preset speed difference and the third preset speed difference; If the speed difference V0 is less than the first preset speed difference, the first speed correction coefficient m1 is selected to correct the risk value F, and the corrected value is F×m1; If the speed difference V0 is greater than or equal to the first preset speed difference, and the speed difference V0 is less than the second preset speed difference, then the second speed correction coefficient m2 is selected to correct the risk value F, and the corrected value is F×m2; If the speed difference V0 is greater than or equal to the second preset speed difference, and the speed difference V0 is less than the third preset speed difference, then the third speed correction coefficient m3 is selected to correct the risk value F, and the corrected value is F×m3; If the speed difference V0 is greater than or equal to the third preset speed difference, the fourth speed correction coefficient m4 is selected to correct the risk value F, and the corrected value is F×m4; wherein 1<m1<m2<m3<m4<1.

2.

6. The shell state recognition method based on local optimization and support vector machine according to claim 5 is characterized in that: The risk value is adjusted according to the depth difference between the diving depth and the depth threshold to obtain a final risk value of the shell, including: Determine the depth difference H0 between the diving depth H and the depth threshold Hm, H0 = H-Hm; If the driving speed is less than or equal to the speed threshold, the risk value F is adjusted according to the depth difference H0 to obtain a final risk value of the shell; If the driving speed is greater than the speed threshold, the corrected F×mi is adjusted according to the depth difference H0, i=1, 2, 3, 4, to obtain the final risk value of the shell.

7. The shell state recognition method based on local optimization and support vector machine according to claim 6 is characterized in that: The risk value F is adjusted according to the depth difference H0 to obtain a final risk value of the shell, including: Presetting a first preset depth difference, a second preset depth difference and a third preset depth difference, wherein the first preset depth difference, the second preset depth difference and the third preset depth difference increase in sequence; adjusting the risk value F according to the relationship between the depth difference and the first preset depth difference, the second preset depth difference and the third preset depth difference; If the depth difference H0 is less than the first preset depth difference, a first depth correction coefficient n1 is selected to correct the risk value F, so that a final risk value is obtained as F×n1; If the depth difference H0 is greater than or equal to the first preset depth difference, and the depth difference H0 is less than the second preset depth difference, then a second depth correction coefficient n2 is selected to correct the risk value F, so that a final risk value is obtained as F×n2; If the depth difference H0 is greater than or equal to the second preset depth difference, and the depth difference H0 is less than the third preset depth difference, a third depth correction coefficient n3 is selected to correct the risk value F, so that a final risk value is obtained as F×n3; If the depth difference H0 is greater than or equal to the third preset depth difference, the fourth depth correction coefficient n4 is selected to correct the risk value F to obtain a final risk value of F×n4; wherein 1<n1<n2<n3<n4<1.

2.

8. The shell state recognition method based on local optimization and support vector machine according to claim 7 is characterized in that: The corrected F×mi is adjusted according to the depth difference H0 to obtain the final risk value of the shell, including: adjusting the corrected F×mi according to the relationship between the depth difference and the first preset depth difference, the second preset depth difference and the third preset depth difference; If the depth difference H0 is less than the first preset depth difference, the fifth depth correction coefficient n5 is selected to correct the corrected F×mi to obtain a final risk value of F×mi×n5; If the depth difference H0 is greater than or equal to the first preset depth difference, and the depth difference H0 is less than the second preset depth difference, a sixth depth correction coefficient n6 is selected to correct the risk value F, and a final risk value is obtained as F×mi×n6; If the depth difference H0 is greater than or equal to the second preset depth difference, and the depth difference H0 is less than the third preset depth difference, then the seventh depth correction coefficient n7 is selected to correct the risk value F, and a final risk value is obtained as F×mi×n7; If the depth difference H0 is greater than or equal to the third preset depth difference, the eighth depth correction coefficient n8 is selected to correct the risk value F to obtain a final risk value of F×mi×n8; wherein 1<n5<n6<n7<n8<1.2, and n1<n5, n2<n6, n3<n7, n4<n8.

9. The shell state recognition method based on local optimization and support vector machine according to claim 1, characterized in that: Generating a driving strategy for the underwater vehicle according to the final risk value includes: Presetting a first preset final risk value and a second preset final risk value, wherein the first preset final risk value is smaller than the second preset final risk value; When the final risk value is less than the first preset final risk value, the driving strategy of the underwater vehicle is generated to be that the underwater vehicle can continue to drive; When the final risk value is greater than or equal to the first preset final risk value, and the final risk value is less than the second preset final risk value, a driving strategy of the underwater vehicle is generated that the underwater vehicle can continue to drive, but the driving time does not exceed the preset time; When the final risk value is greater than or equal to the second preset final risk value, the driving strategy of the underwater vehicle is generated to stop driving and return immediately.

Images