Optical fiber angular velocity sensor convenient to assemble

Abstract

The invention relates to the technical field of sensors, in particular to an optical fiber angular velocity sensor convenient to assemble, which comprises a sensor data collection module used for collecting sensing data of the sensor; the data preprocessing module is used for preprocessing the sensing data to obtain sensor preprocessing data; the whale optimization algorithm module is used for iteratively optimizing the design parameters of the sensor to obtain optimal design parameters; the model training module is used for establishing a sensor performance prediction model by using a deep belief network and predicting performance indexes of the sensor according to the optimal design parameters; and the result verification module is used for manufacturing a sensor sample according to the optimized design parameters, carrying out performance test, judging whether a performance test result is within a preset range or not, and outputting an optimal performance index if the performance test result is within the preset range. The whale optimization algorithm and the deep belief network are used for obtaining the optimal design parameters, the universality and robustness of the sensor are enhanced, and the sensor can better adapt to various complex environments.

Description

Technical Field

[0001] The invention relates to the technical field of sensors, and in particular to an optical fiber angular velocity sensor which is easy to assemble. Background Art

[0002] The fiber optic angular velocity sensor is a high-precision angular velocity measurement device based on optical principles. The working principle of the fiber optic angular velocity sensor is based on the Sagnac effect. When light enters the fiber optic loop from the light source, it will be reflected multiple times in the loop and form an interference loop.

[0003] Existing theoretical models cannot fully and accurately describe the behavior of sensors in practical applications. Parameters such as the refractive index, loss, and nonlinear effects of optical fibers may be affected by a variety of factors in the actual environment, resulting in deviations between theoretical predictions and actual performance. The properties of optical fiber materials, such as the thermal expansion coefficient and the temperature coefficient of refractive index, may vary due to slight differences in the manufacturing process. These changes may cause unstable sensor performance, making the predicted results inconsistent with the actual performance. Fiber optic angular velocity sensors are sensitive to changes in environmental factors such as temperature, humidity, and pressure. These changes may cause unstable sensor performance or even failure.

[0004] The present invention provides an optical fiber angular velocity sensor that is easy to assemble and can efficiently search the parameter space of the sensor structure to find the optimal or approximately optimal design solution, so that the designer can have a more comprehensive understanding of the performance of the sensor in different environments and make more reasonable decisions. Summary of the invention

[0005] The present invention provides an optical fiber angular velocity sensor which is easy to assemble, so as to solve the defect of inaccurate performance prediction of the optical fiber angular velocity sensor in the prior art.

[0006] The present invention provides an optical fiber angular velocity sensor which is easy to assemble, comprising: a sensor data collection module, which is used to collect sensor perception data, wherein the perception data includes design parameters and performance indicators.

[0007] The data preprocessing module is used to preprocess the perception data to obtain sensor preprocessing data.

[0008] The whale optimization algorithm module is used to obtain the most suitable design parameter combination by iteratively optimizing the design parameters of the sensor.

[0009] The model training module uses a deep belief network to establish a sensor performance prediction model and predicts the performance indicators of the sensor based on the optimal design parameter combination.

[0010] The result verification module makes sensor samples according to the optimized design parameters and performs performance tests to determine whether the performance test results are within the preset range. If so, the optimal performance index is output.

[0011] According to the optical fiber angular velocity sensor that is easy to assemble provided by the present invention, in the sensor data collection module, the design parameters and performance indicators include: Design parameters: shape, size, material properties, stress distribution, thermal deformation, strength.

[0012] Performance indicators: stability, resolution, accuracy, power consumption.

[0013] According to the optical fiber angular velocity sensor that is easy to assemble provided by the present invention, the sensor preprocessing data includes: Data cleaning is used to clean the perception data and remove duplicate, erroneous or abnormal data points.

[0014] Feature selection is used to extract features that affect sensor performance from perception data.

[0015] Data normalization is used to normalize design parameters and performance indicators to the same scale.

[0016] According to the fiber optic angular velocity sensor that is easy to assemble provided by the present invention, the optimization process of the whale optimization algorithm module includes: Set the number of design parameter combinations to N and the maximum number of iterations to T max , initialize the position of each parameter combination in the design parameter combination, denoted as X i (i=1,2,3,…,n).

[0017] Evaluate the fitness value of each parameter combination, and select n parameter individuals with the best fitness value as the initial population position, record the optimal fitness value and its corresponding position at this time, and update the parameters a, A, C, I, and p at the same time.

[0018] The position of the parameter combination is updated using the formula to obtain the position update stage.

[0019] Determine whether the algorithm termination condition is met. If so, output the optimal target value. Otherwise, repeat the position update phase until the termination condition is met.

[0020] According to the optical fiber angular velocity sensor that is easy to assemble provided by the present invention, the position update phase includes: If p is less than 0.5 and |A| is less than 1, the individual position of the updated parameter combination is position one.

[0021] If p is less than 0.5 and |A| is greater than or equal to 1, the individual position of the updated parameter combination is position two.

[0022] If p is greater than or equal to 0.5, the individual position of the updated parameter combination is position three.

[0023] Recalculate the fitness value of the updated whale and update the current optimal solution.

[0024] According to the optical fiber angular velocity sensor that is easy to assemble provided by the present invention, in the position update phase, the update formula of position one is:

[0025] In the formula, p is the probability of state selection, t is the current number of iterations, A and C are coefficients, and X * (t) is the design parameter combination of a group of sensors with the best fitness at present, X(t) represents the design parameter combination in the current iteration, D is the distance between the current design parameter combination and the target parameter combination, I is a random value, and b is the parameter of the exponential function.

[0026] According to the optical fiber angular velocity sensor that is easy to assemble provided by the present invention, in the position update phase, the update formula of position two is:

[0027] In the formula, X rand is the location of random design parameter combinations.

[0028] According to the optical fiber angular velocity sensor that is easy to assemble provided by the present invention, the model training module includes: The model building unit is used to build a model according to the optimal design parameter combination and the perception data.

[0029] The model training unit uses a combination of unsupervised learning and supervised learning to train the DBN model.

[0030] The model evaluation unit is used to evaluate the prediction performance of the model using the test set and compare the performance differences between different models.

[0031] According to the optical fiber angular velocity sensor that is easy to assemble provided by the present invention, the model building unit includes: The network structure of DBN is designed according to the optimal design parameter combination and the perception data. The network structure includes the number of nodes and connection mode of the input layer, hidden layer and output layer.

[0032] The number of input layer nodes is the number of design parameter combinations.

[0033] The number of nodes in the output layer is the number of performance indicators.

[0034] The number of hidden layer nodes is determined by historical sensor data.

[0035] The optical fiber angular velocity sensor that is easy to assemble according to the present invention includes: a model training unit includes: The unsupervised pre-training subunit is used to train the network layer by layer using an unsupervised learning method. The first layer of RBM is trained first, and its learned features are used as the input of the second layer of RBM. The second layer of RBM is then trained until all layers are trained.

[0036] Supervised fine-tuning subunits,After the unsupervised pre-training is completed, DBN uses a supervised learning method to,finetune the entire network, using the back-propagation algorithm to adjust the,weights and bias terms of the network to minimize the prediction error.

[0037] The fiber optic angular velocity sensor provided by the present invention is easy to assemble. The design parameters of the sensor are iteratively optimized by the whale optimization algorithm to obtain the optimal design parameters. A sensor performance prediction model is established by using a deep belief network. The performance indicators of the sensor are predicted according to the optimal design parameters, which solves the problem of inaccurate performance prediction of the fiber optic angular velocity sensor. The beneficial effects achieved are as follows: Traditional sensor design methods often rely on a lot of experiments and trial and error. Using the whale optimization algorithm for simulation and simulation can greatly reduce the number of experiments, reduce design costs and time costs. The whale optimization algorithm has the characteristics of fast convergence and can find an approximate optimal solution in a short time. This allows designers to quickly iterate and optimize sensor structures and improve design efficiency.

[0038] The whale optimization algorithm can fully simulate the working state of the sensor under different assembly conditions, so as to more accurately predict the actual performance of the sensor. Through simulation and simulation, potential performance problems can be identified and solved in the design stage, thereby reducing the number and cost of experimental verification. The algorithm can automatically search for the optimal combination of structural parameters so that the sensor can achieve the best balance in stress, thermal deformation, strength, etc., thereby improving the overall performance. The whale optimization algorithm can automatically search for the optimal solution in the sensor structural parameter space without manual attempts one by one, which greatly improves the design efficiency. Through simulation and simulation, the effects of different design schemes can be quickly evaluated, thereby accelerating the design iteration process and shortening the overall design cycle. The number and complexity of experimental verification are reduced, and the material consumption and labor costs in the design process are reduced. This helps to evaluate the adaptability and stability of sensors in different environments. Through the optimization algorithm, sensor structural parameters that perform well under different conditions can be found. This enhances the versatility and robustness of the sensor, enabling it to better adapt to various complex environments.

[0039] DBN can automatically learn meaningful feature representations from sensor data, which are critical for subsequent performance prediction tasks. It avoids the uncertainty and difficulty of manual feature extraction and improves the accuracy and robustness of the prediction model. DBN can model the data generation process. In sensor performance prediction, this helps to understand the distribution and change patterns of sensor data and provide a more accurate basis for prediction. DBN can learn and understand the distribution of data and generate new data samples based on the learned model. This can simulate different scenarios and conditions when predicting sensor performance, and further verify and optimize the prediction model.

[0040] WOA can efficiently search for the optimal parameter configuration of DBN, such as the number of layers of the network, the number of nodes in each layer, the learning rate, etc. This optimization process can significantly improve the performance of DBN, making it more accurate and efficient when processing complex tasks. By optimizing the parameters of DBN through WOA, a set of optimal parameter configurations can be obtained, which not only performs well on the current data set, but also maintains good performance on unknown data sets. After DBN is combined with WOA, the generalization ability of the model is enhanced, and it can better adapt to different application scenarios and data distributions. The traditional DBN training process may involve a lot of parameter adjustments and experiments to find the best model configuration. As an efficient optimization algorithm, WOA can significantly reduce this trial and error process and accelerate model training. By quickly converging to the optimal solution, WOA enables DBN to reach a stable performance level faster, thereby saving training time and computing resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 It is a module diagram of an optical fiber angular velocity sensor that is easy to assemble provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0044] Combine the following Figure 1 The fiber optic angular velocity sensor of the present invention which is easy to assemble is described.

[0045] like Figure 1 As shown, the optical fiber angular velocity sensor provided by the embodiment of the present invention is easy to assemble and includes: an optical fiber coil, a laser, a beam splitter, a signal detection system, and auxiliary components.

[0046] Fiber optic coil: The fiber optic coil consists of multiple turns of single-pass optical fiber wound on a skeleton. When the system has an angular velocity relative to the earth's inertial system, the two beams of light propagating clockwise and counterclockwise in the fiber optic coil will produce an optical path difference or phase difference, which is the Sagnac effect. The fiber optic connector is a standardized interface that provides mechanical stability and optical alignment to ensure low loss of optical signals during transmission. These connections are usually achieved through fiber optic connectors or fiber optic adapters to ensure efficient transmission of optical signals.

[0047] Laser: used to emit a collimated light beam, which is split into two beams of equal intensity after passing through a beam splitter. The performance of the laser directly affects the measurement accuracy and stability of the sensor.

[0048] Beam splitter: A beam splitter is usually an optical element, such as a semi-transparent mirror or a fiber coupler, which is used to split the beam emitted by the laser into two beams of equal intensity and guide them to the two ends of the optical fiber coil respectively so as to propagate in the clockwise and counterclockwise directions in the optical fiber coil.

[0049] Focusing lens: By placing the lens near the end of the optical fiber coil and adjusting the position and focal length of the lens, the focusing lens focuses the two split beams of light onto the end face of the optical fiber coil to ensure that the light can efficiently enter the optical fiber coil and propagate along it. The focusing lens can improve the efficiency of the optical signal entering the optical fiber coil and reduce the loss and reflection of the optical signal.

[0050] Signal detection system: used to detect the phase difference or optical path difference between two beams of light propagating from the optical fiber coil, and convert it into an electrical signal for further processing and analysis. The signal detection system usually includes components such as photodetectors and signal processing circuits. The signal detection system includes components such as photodetectors and signal processing circuits. Photodetectors are used to convert optical signals into electrical signals, while signal processing circuits are used to amplify, filter and digitize electrical signals for further analysis and calculation.

[0051] Other auxiliary components: Fiber optic angular velocity sensors also include other auxiliary components such as fiber optic connectors, adapters, and housings. Auxiliary components are used to protect the optical and electronic components inside the sensor from interference and damage from the external environment, making it easier to install and integrate into larger systems or equipment.

[0052] The laser is connected to the beam splitter, and the two ends of the optical fiber coil are connected to the two output ports of the beam splitter respectively. After the two beams of light propagate in the optical fiber coil, they will return to the beam splitter again, and the optical signals after interference here will be guided to the signal detection system. The optical fiber connector is electrically connected to the adapter, and the housing is threadedly connected to the front connector.

[0053] The sensor data collection module is used to collect the sensor's perception data, which includes design parameters and performance indicators.

[0054] Sensor performance prediction and optimization aims to improve the performance indicators of sensors by adjusting their design parameters, such as shape, size, material properties, stress distribution, thermal deformation, and strength. The performance indicators include stability, resolution, accuracy, power consumption, etc.

[0055] The data preprocessing module is used to preprocess the sensor perception data to obtain sensor preprocessing data. Sensor preprocessing data includes: data cleaning, feature selection, and data normalization.

[0056] Data cleaning is used to clean the collected perception data and remove duplicate, erroneous or abnormal data points.

[0057] Feature selection: Extracting features that have a significant impact on sensor performance from raw data, which helps reduce model complexity and improve prediction accuracy.

[0058] Data normalization: Normalizing or standardizing design parameters and performance metrics to ensure they are on the same scale helps machine learning algorithms learn and predict better.

[0059] The whale optimization algorithm optimization module obtains the most suitable design parameter combination by iteratively optimizing the design parameters of the sensor.

[0060] The Whale Optimization Algorithm is used to simulate and model the structure of the sensor to predict and optimize its performance. The optimal structural design solution is found by simulating the stress distribution, thermal deformation, strength, stability, accuracy and resolution under different assembly conditions.

[0061] The whale optimization algorithm uses random or optimized search agents to simulate the hunting strategies of whales in nature, thereby realizing the exploration and optimization of the search space. The spiral model is used to simulate the attack mechanism of whales, which enables the algorithm to explore the solution space more flexibly and comprehensively during the search process. Different positions of the whale population during the hunting process represent different feasible solutions. This is like the process of finding the optimal solution in the solution space. Each iteration is like a whale swimming one step in the ocean, and by constantly adjusting its position, it finally finds the global optimal solution. There are three search methods in this algorithm, namely, encircling prey, hunting with bubble nets, and searching for prey.

[0062] In the early stage of encircling prey, the design parameters of a set of sensors represent a feasible solution. Since the position of the optimal individual is unknown in the search space, the position of the prey needs to be determined first in order to encircle it. The whale optimization algorithm takes the current best potential solution as the target object and adopts the encirclement strategy to update its search position.

[0063] The hunting formula is expressed as:

[0064] In the formula, t is the current iteration number, A and C are coefficients, and X * (t) is the design parameter of a set of sensors with the best fitness at present, X(t) represents the design parameter in the current iteration, and D is the distance between the current design parameter and the target parameter. If there is a better candidate solution, the value of X*(t) is updated in each iteration. The specific calculation formulas of A and C are expressed as:

[0065] In the formula, r1 and r2 are random values ​​in the interval (0, 1), and in the entire iteration process, the value of parameter a will linearly decrease from 2 to 0. This change process runs through the exploration phase and the utilization phase of the algorithm. t represents the current number of iterations, T max The maximum number of iterations for the initial setting.

[0066] During the hunting phase, whales capture prey by reducing the encirclement and releasing bubbles in a spiral rotation to form a bubble net. That is, when humpback whales perform spiral motion, they will simultaneously reduce the encirclement to ensure that they can effectively shrink and surround the prey. According to the calculation formula of A, A at this time is a value randomly selected in the interval [-a, a]. When the random value range of A is limited to [-1, 1], the updated design parameter X(t+1) will fluctuate between the current design parameter X(t) and the optimal design parameter X*(t), thereby achieving effective encirclement of the target parameter. The formula for updating the current design parameters is:

[0067] In the formula, is the distance between the design parameter and the target design parameter, b is a constant, defined as the logarithmic spiral shape presented when surrounding the prey, I is a random value ranging from [-1, 1], and b is the parameter of the exponential function.

[0068] When a whale is surrounding its prey, it is surrounding its prey and spiraling up at the same time. Assuming that the probability of state selection between the surrounding mechanism and the spiraling up is 0.5, the surrounding formula is expressed as:

[0069] In the formula, p is the probability of state selection, and the random range is [0,1]. By adjusting the probability p, parameter a, and fluctuation range A, the design parameters are migrated to better parameters. As the number of iterations increases, the parameter A and the convergence factor a in the above formula will continue to decrease until |A|<1, indicating that the current local optimal solution is obtained.

[0070] Prey search phase: In order to enable the whale population to search more fully in the solution space, the algorithm updates the position by the distance between different whale individuals to ensure random search. During the search process, individual whales tend to approach randomly selected peers and use their position information to update their own positions. This method actually guides whales away from the current location of prey, thereby improving the algorithm's global optimization ability.

[0071] The mathematical model is expressed as:

[0072] Where D is the distance between the current design parameter and the randomly selected design parameter, X rand is the position of the random design parameter. When |A|≥1, the algorithm will randomly select a design parameter and update the positions of other design parameters accordingly. This is to increase the randomness of the search and help find better quality prey.

[0073] When the maximum number of iterations is reached or the termination condition is met, the algorithm iteration stops and the current optimal solution, that is, the optimal sensor design solution, is output.

[0074] The model training module uses a deep belief network to establish a sensor performance prediction model and predict the performance indicators of the sensor based on the optimal design parameters.

[0075] The model building unit is used to build a model according to the optimal design parameter combination and the perception data.

[0076] Network structure design, design the network structure of DBN according to the optimal design parameters and perception data, including the number of nodes and connection methods of the input layer, hidden layer and output layer. The number of input layer nodes corresponds to the number of design parameters, the number of output layer nodes corresponds to the number of performance indicators, and the number of hidden layers and nodes needs to be analyzed and determined based on experiments to find the optimal network structure.

[0077] The model training unit uses a combination of unsupervised learning and supervised learning to train the DBN model.

[0078] Unsupervised pre-training uses unsupervised learning methods to train the network layer by layer. First, train the first layer of RBM, use the learned features as the input of the second layer of RBM, then train the second layer of RBM, and so on, until all layers are trained. This process enables each layer of RBM to learn an effective representation of the input data. Use the contrastive divergence algorithm to train the RBM layer by layer to ensure that the feature information is preserved as much as possible when the feature vector is mapped to different feature spaces.

[0079] Supervised fine-tuning: After unsupervised pre-training is completed, DBN uses supervised learning methods to fine-tune the entire network. This stage usually uses the back-propagation algorithm to adjust the weights and biases of the network to minimize the prediction error. Through fine-tuning, DBN can further improve its performance on specific tasks. Use the back-propagation algorithm for global parameter optimization to adjust the weights and biases of the entire network to minimize the prediction error.

[0080] The model evaluation unit is used to evaluate the prediction performance of the model using the test set and compare the performance differences between different models. The model is evaluated using techniques such as cross-validation to select the best hyperparameter combination.

[0081] The result verification module makes sensor samples according to the optimized design parameters and performs performance tests to determine whether the performance test results are within the preset range. If so, the optimal performance index is output.

[0082] The fiber optic angular velocity sensor provided by the present invention is easy to assemble. The design parameters of the sensor are iteratively optimized by the whale optimization algorithm to obtain the optimal design parameters. A sensor performance prediction model is established by using a deep belief network. The performance indicators of the sensor are predicted according to the optimal design parameters, which solves the problem of inaccurate performance prediction of the fiber optic angular velocity sensor. The beneficial effects achieved are as follows: WOA can efficiently search for the optimal parameter configuration of DBN, such as the number of layers of the network, the number of nodes in each layer, the learning rate, etc. This optimization process can significantly improve the performance of DBN, making it more accurate and efficient when processing complex tasks. By optimizing the parameters of DBN through WOA, a set of optimal parameter configurations can be obtained, which not only performs well on the current data set, but also maintains good performance on unknown data sets. This means that after DBN is combined with WOA, the generalization ability of the model is enhanced and can better adapt to different application scenarios and data distributions. The traditional DBN training process may involve a lot of parameter adjustments and experiments to find the best model configuration. As an efficient optimization algorithm, WOA can significantly reduce this trial and error process and accelerate model training. By quickly converging to the optimal solution, WOA enables DBN to reach a stable performance level faster, thereby saving training time and computing resources.

[0083] The device embodiments described above are merely illustrative, wherein the units described as separate components are or are not physically separated, and the components shown as units are or are not physical units, that is, located in one place, or distributed on multiple network units. Some or all of the modules are selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative work.

[0084] Through the description of the above implementation modes, those skilled in the art clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, and of course by hardware. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art is embodied in the form of a software product, and the computer software product is stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for a computer device (a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments are still modified, or some of the technical features thereof are replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An optical fiber angular velocity sensor that is easy to assemble, characterized in that: include: A sensor data collection module, used to collect sensor perception data, wherein the perception data includes design parameters and performance indicators; A data preprocessing module, used to preprocess the sensing data to obtain sensor preprocessing data; Whale optimization algorithm module, used to iteratively optimize the design parameters of the sensor to obtain the most suitable design parameter combination; A model training module, using a deep belief network to establish a sensor performance prediction model, and predicting the performance index of the sensor according to the optimal design parameter combination; The result verification module makes sensor samples according to the optimized design parameters and performs performance tests to determine whether the performance test results are within the preset range. If so, the optimal performance index is output.

2. The optical fiber angular velocity sensor that is easy to assemble according to claim 1, characterized in that: In the sensor data collection module, the design parameters and performance indicators include: Design parameters: shape, size, material properties, stress distribution, thermal deformation, strength; Performance indicators: stability, resolution, accuracy, power consumption.

3. The optical fiber angular velocity sensor that is easy to assemble according to claim 1, characterized in that: Sensor preprocessing data includes: Data cleaning, used to clean the perception data and remove duplicate, erroneous or abnormal data points; Feature selection, used to extract features that affect sensor performance from perception data; Data normalization is used to normalize design parameters and performance indicators to the same scale.

4. The optical fiber angular velocity sensor that is easy to assemble according to claim 1, characterized in that: The optimization process in the whale optimization algorithm module includes: Set the number of design parameter combinations to N and the maximum number of iterations to T max , initialize the position of each parameter combination in the design parameter combination, denoted as X i (i=1,2,3,…,n); Evaluate the fitness value of each parameter combination, and select n parameter individuals with the best fitness value as the initial population position, record the optimal fitness value and its corresponding position at this time, and update the parameters a, A, C, I, and p at the same time; The position of the parameter combination is updated using the formula to obtain the position update stage; Determine whether the algorithm termination condition is met. If so, output the optimal target value. Otherwise, repeat the position update phase until the termination condition is met.

5. The optical fiber angular velocity sensor that is easy to assemble according to claim 4, characterized in that: The location update phase includes: If p is less than 0.5 and |A| is less than 1, the individual position of the updated parameter combination is position one; If p is less than 0.5 and |A| is greater than or equal to 1, the individual position of the updated parameter combination is position 2; If p is greater than or equal to 0.5, the individual position of the updated parameter combination is position three; Recalculate the fitness value of the updated whale and update the current optimal solution.

6. The optical fiber angular velocity sensor that is easy to assemble according to claim 5, characterized in that: In the position update phase, the update formula for position one is: ； In the formula, p is the probability of state selection, t is the current number of iterations, I is the random value, A and C are coefficients, and X * (t) is the design parameter combination of a group of sensors with the best fitness at present, X(t) represents the design parameter combination in the current iteration, D is the distance between the current design parameter combination and the target parameter combination, and b is the parameter of the exponential function.

7. The optical fiber angular velocity sensor that is easy to assemble according to claim 6, characterized in that: In the position update phase, the update formula for position 2 is: ； ； Where, X rand is the location of random design parameter combinations.

8. The optical fiber angular velocity sensor that is easy to assemble according to claim 1, characterized in that: The model training module includes: A model building unit, used for building a model according to the optimal design parameter combination and the perception data; Model training unit, which trains DBN model by combining unsupervised learning and supervised learning; The model evaluation unit is used to evaluate the prediction performance of the model using the test set and compare the performance differences between different models.

9. The optical fiber angular velocity sensor that is easy to assemble according to claim 8, characterized in that: The model building unit includes: Designing a network structure of a DBN according to the optimal design parameter combination and the perception data, wherein the network structure includes the number of nodes and connection mode of an input layer, a hidden layer, and an output layer; The number of nodes in the input layer is the number of design parameter combinations; The number of nodes in the output layer is the number of performance indicators; The number of hidden layer nodes is determined by historical sensor data.

10. The optical fiber angular velocity sensor that is easy to assemble according to claim 8, characterized in that: The model training unit includes: The unsupervised pre-training subunit is used to train the network layer by layer using an unsupervised learning method. The first layer of RBM is trained first, and its learned features are used as the input of the second layer of RBM. The second layer of RBM is then trained until all layers are trained. Supervised fine-tuning subunits,After the unsupervised pre-training is completed, DBN uses a supervised learning method to,finetune the entire network, using the back-propagation algorithm to adjust the,weights and bias terms of the network to minimize the prediction error.

Images