Computer equipment positioning method based on mathematical model

Through the computer equipment positioning method based on mathematical model, using statistics and neural network technology to process original data, generate and match positioning models, the problem of equipment being unable to quickly locate when satellite positioning is lost, and efficient in a signal circuit breaker environment is achieved Positioning and precise position prediction.

CN118936475AInactive Publication Date: 2024-11-12FUJIAN POST & TELECOMM SCHOOL
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411060807.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-05
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

When the satellite positioning system is out of contact, the traditional positioning method based on satellite signals cannot work properly, resulting in the equipment being unable to quickly determine the position.

Method used

Using a computer equipment positioning method based on mathematical model, collects original data through the acquisition unit, preprocesses and trains the mathematical model, generates a mobile positioning model, and uses statistical methods or neural network deep learning technology to match and predict, providing missing devices Pre-determined position.

Benefits of technology

In a signal-breaking environment, the predefined positioning position of the equipment is effectively obtained, the model matching error is reduced, the various environments are adapted to real-time positioning guidance, and the positioning accuracy and reliability are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118936475A_ABST
    Figure CN118936475A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of positioning methods, and particularly relates to a mathematical model-based computer equipment positioning method, which is characterized in that a computer equipment positioning system is included, and the computer equipment positioning system acquires original data of test equipment through an acquisition unit; according to the computer equipment positioning system, the collected original data is extracted into the mathematical model database through the collected information preprocessing unit, in the scheme, the preset positioning position of the lost equipment can be effectively obtained through the correspondingly-arranged computer equipment positioning system, the preset route is effectively provided for the lost equipment through model prediction, and the positioning accuracy of the lost equipment is improved. The device is widely applied to positioning of a signal disconnection position, effectively adapts to various applicable environments by utilizing various training modes of equipment, and effectively plays a positioning and guiding role.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of positioning methods, and in particular relates to a computer equipment positioning method based on a mathematical model. Background Art

[0002] In some cases, the satellite positioning system may fail, causing the device to be unable to determine its own location. In this case, the traditional positioning method based on satellite signals may not work properly. Therefore, the purpose of the present invention is to improve the ability of device location estimation in an environment where satellite positioning is disconnected, so as to ensure that users can obtain accurate device location information in a timely manner. In the prior art, a positioning communication device, a positioning method and a computer storage medium with reference patent number: 201810404620X adopts a positioning algorithm based on the signal arrival angle (Phase Difference of Arrival, PDOA) for positioning. Under ideal circumstances, a higher positioning accuracy can be achieved. At the same time, the positioning signals received by the anchor node and the beacon can determine the orientation information of the beacon relative to itself. In this way, when the predetermined beacon signal is lost, it is impossible to quickly locate according to the anchor node in time. For this reason, we propose a computer device positioning method based on a mathematical model. Summary of the invention

[0003] The purpose of the present invention is to provide a computer equipment positioning method based on a mathematical model to solve the problems raised in the above background technology.

[0004] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a computer equipment positioning method based on mathematical model, characterized in that: it includes a computer equipment positioning system, the computer equipment positioning system collects the original data of the test equipment through the collection unit, the computer equipment positioning system extracts the collected original data into the mathematical model database through the collection information preprocessing unit, the computer equipment positioning system performs model training on the collected data in the mathematical model data through the model training unit to generate a mobile positioning model, the computer equipment positioning system matches the mobile positioning model with the corresponding predetermined positioning model in the mathematical model database through the matching unit, and displays the predetermined positioning model through the external display unit.

[0005] Furthermore, when the matching unit measures and matches the predetermined positioning model, the matching degree is monitored by the test and calibration unit. When the acquisition unit acquires new raw data provided by the test device, the closest predetermined positioning model is adjusted accordingly.

[0006] Furthermore, the acquisition unit acquires raw data of the test device including moving speed, direction and timestamp.

[0007] Furthermore, the computer algorithms selected in the mathematical model library for learning and training the predetermined positioning model include statistical methods or neural network deep learning technology.

[0008] Furthermore, the statistical method trains the mathematical model by preprocessing the original data of the fixed section acquisition unit and obtaining the predetermined route band map through linear embedding, thereby constructing a predetermined positioning model, and migrating the existing model knowledge into the mathematical model library in the predetermined positioning model.

[0009] Furthermore, the neural network deep learning technology trains the mathematical model by preprocessing the raw data of the fixed section acquisition unit to select a predetermined neural network structure to build a neural network model, sets a loss function and an optimizer, and trains the neural network model by defining the loss function and evaluation indicators, and regularly evaluates the performance through adjustable hyperparameters to form a predetermined positioning model.

[0010] Furthermore, the neural network deep learning technology uses real-time variables to perform data enhancement on images in the database when training the mathematical model, and uses regularization strategies to reduce redundant parameters.

[0011] Furthermore, a real-time positioning algorithm is introduced into the matching unit.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. In this solution, the predetermined location of the lost device can be effectively obtained through the corresponding computer equipment positioning system, and the model prediction is used to effectively provide the lost device with a predetermined route, which is widely used in the positioning of signal break locations. At the same time, various training methods of the equipment are used to effectively respond to various applicable environments and effectively play a positioning guidance effect.

[0013] 2. Through the corresponding update system set up in the back end, the update operation can be matched in real time to reduce the model matching error. At the same time, model training can effectively increase the trajectory overlap with specific physical devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0015] Figure 1 It is a schematic diagram of the overall process structure of the present invention. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0017] In the description of the present invention, it is necessary to understand that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0018] Reference Figure 1 The present invention proposes a technical solution: a computer device positioning method based on a mathematical model, characterized in that: it includes a computer device positioning system, the computer device positioning system collects the original data of the test device through a collection unit to ensure the accuracy of the data: when collecting data such as the moving speed, direction and timestamp of the device, the accuracy and integrity of the data must be ensured. Errors can be reduced by regularly checking the device status, updating the firmware, etc.

[0019] Increase data diversity: Collect a variety of information related to the device's movement speed, direction, and timestamp data, such as environmental factors such as temperature, humidity, and air pressure. This data can help you better understand device behavior and thus improve the accuracy of path planning.

[0020] The computer equipment positioning system extracts the collected raw data into the mathematical model database through the information collection preprocessing unit. In the preprocessing process, anomaly detection algorithms are used to detect outliers or abnormal timestamps, which may be caused by equipment failure, data transmission problems, etc. Removing such abnormal data will help improve the quality of subsequent analysis.

[0021] By smoothing the data, noise can be eliminated and the complexity of the multi-road condition prediction model can be reduced. You can try to use methods such as sliding average and exponentially weighted moving average for smoothing.

[0022] The computer equipment positioning system performs model training on the collected data in the mathematical model data through the model training unit to generate a mobile positioning model. The computer equipment positioning system matches the mobile positioning model with the corresponding predetermined positioning model in the mathematical model database through the matching unit, and displays the predetermined positioning model through the external display unit.

[0023] When the matching unit measures and matches the predetermined positioning model, the matching degree is monitored through the test verification unit. When the acquisition unit collects new raw data provided by the test equipment, it will adjust the closest predetermined positioning model accordingly, and introduce a real-time positioning algorithm into the matching unit.

[0024] During the database model deployment phase, the model can be updated in real time. Specifically, with the launch of new devices, software updates, and other factors, the model should be updated regularly to ensure its adaptability to new data.

[0025] Increase scalability by designing an extensible architecture to allow the addition of new sensors, data sources, or model structures in the future.

[0026] Ensure security, specifically considering data security and privacy protection issues, such as using encryption technology to protect user data.

[0027] The acquisition unit collects the original data of the test equipment including the moving speed, direction and timestamp. The computer algorithm including statistical method or neural network deep learning technology is selected for learning and training the predetermined positioning model in the mathematical model library. In this scheme, the predetermined positioning position of the lost device can be effectively obtained through the corresponding computer equipment positioning system. The model prediction is used to effectively provide the lost device with a predetermined route, which is widely used in the positioning of the signal break position. At the same time, the equipment's various training methods are used to effectively cope with various applicable environments and effectively play a positioning guidance effect. Embodiment 1

[0028] The statistical method is used to train the mathematical model by preprocessing the raw data of the fixed section acquisition unit and obtaining the predetermined route band map through linear embedding, so as to construct a predetermined positioning model, in which the existing model knowledge is transferred to the mathematical model library.

[0029] The preprocessing steps in the above steps mainly perform principal component analysis through PCA to extract predetermined elements and delete corresponding irrelevant characteristics such as outliers, fluctuating road sections, weather factors, etc.

[0030] Among them, PCA may not be able to completely solve all data dimensionality reduction problems. In some cases, it may be necessary to use other dimensionality reduction methods, such as t-SNE, UMAP, etc. These methods can reduce the computational complexity of the model while retaining more effective information; For some features, such as timestamp, temperature, etc., their dimensionality reduction may lead to information loss. In this case, it is possible to consider fusing these features with other features to achieve information complementarity. Embodiment 2

[0031] The neural network deep learning technology trains the mathematical model by preprocessing the raw data of the fixed section collection unit, selecting a predetermined neural network structure to build the neural network model, setting the loss function and optimizer, and training the neural network model by defining the loss function and evaluation index, and regularly evaluating the performance through adjustable hyperparameters to form a predetermined positioning model. When training the mathematical model, the neural network deep learning technology uses real-time variables to enhance the data of the images in the database, and uses regularization strategies to reduce redundant parameters; The predetermined neural network structure algorithm can use convolutional neural network (CNN) or recurrent neural network (RNN), and the defined loss function and evaluation index, among which the mean square error (MSE) is calculated by calculating the average of the square of the difference between the predicted value and the actual observed value, and the specific value is the same as the square of the unit of the original observed value. It represents the average size of the difference between the model predicted value and the actual observed value. A smaller calculated mean square error (MSE) indicates that the difference between the model's predicted value and the actual observed value is smaller, that is, the model has a better degree of fit.

[0032] The mean absolute error (MAE) is calculated by averaging the absolute values ​​of the differences between the predicted values ​​and the actual observed values. Its specific value has the same unit as the original observed value. It represents the average size of the difference between the model's predicted values ​​and the actual observed values. A smaller mean absolute error (MAE) indicates that the difference between the model's predicted values ​​and the actual observed values ​​is smaller, that is, the model has a better fit.

[0033] The root mean square error (RMSE) is obtained by calculating the mean of the squares of the differences between the predicted values ​​and the actual observed values ​​and taking the square root. Its specific value has the same unit as the original observed value. It can measure the average size of the model's prediction error. A smaller root mean square error (RMSE) indicates that the difference between the model's predicted values ​​and the actual observed values ​​is smaller, that is, the model has a better fit.

[0034] Convolutional neural networks (CNNs) are often used for image and video processing tasks. The advantage is that local features and spatial relationships can be effectively extracted, thereby achieving efficient feature extraction and classification. However, for sequence data, CNNs may not be the best choice because they do not consider time order. Therefore, when processing sequence data such as text, we may choose to use recurrent neural networks (RNNs).

[0035] If the goal is to process time series data, then recurrent neural networks (RNNs) learn long-term dependencies in sequence data and capture temporal order information.

[0036] However, for certain environments, when the performance of the two methods is comparable, we can use a hybrid model that combines a convolutional neural network (CNN) with a recurrent neural network (RNN) to achieve better performance. For example, we can use CNN to extract features and then use RNN for further processing and classification.

[0037] Data augmentation and regularization strategies: Use random transformations, such as rotating, scaling, and flipping images, to perform data augmentation on images in the training set. During model training, apply L2 regularization or other regularization strategies to reduce redundant parameters and improve model generalization capabilities. Evaluate model performance on the test set, generate performance reports, and adjust neural network structure, parameters, or hyperparameters such as learning rate based on the evaluation results to improve model performance.

[0038] Cross-validation can also help us better understand model performance and avoid overfitting during training. Specifically, the data set is divided into multiple subsets, and then the model is trained and validated in each subset in turn. By comparing the validation indicators on each subset, the best parameters or model are selected and applied to the entire data set.

[0039] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A computer equipment positioning method based on a mathematical model, characterized in that: The computer device positioning system includes a computer device positioning system, which collects raw data of the test device through a collection unit, extracts the collected raw data into a mathematical model database through a collection information preprocessing unit, performs model training on the collected data in the mathematical model data through a model training unit to generate a mobile positioning model, matches the mobile positioning model with a corresponding predetermined positioning model in the mathematical model database through a matching unit, and displays the predetermined positioning model through an external display unit.

2. A method for positioning computer equipment based on a mathematical model according to claim 1, characterized in that: When the matching unit measures and matches the predetermined positioning model, the matching degree is monitored by the test and calibration unit. When the acquisition unit acquires new raw data provided by the test equipment, the closest predetermined positioning model is adjusted accordingly.

3. The method for positioning a computer device based on a mathematical model according to claim 1, characterized in that: The acquisition unit acquires raw data of the test device including moving speed, direction and time stamp.

4. The method for positioning computer equipment based on mathematical model according to claim 1, characterized in that: The computer algorithms selected for learning and training the predetermined positioning model in the mathematical model library include statistical methods or neural network deep learning technology.

5. A method for positioning computer equipment based on a mathematical model according to claim 4, characterized in that: The statistical method trains the mathematical model by preprocessing the original data of the fixed section acquisition unit and obtaining the predetermined route band map through linear embedding, thereby constructing a predetermined positioning model, and migrating the existing model knowledge into the mathematical model library in the predetermined positioning model.

6. The method for positioning computer equipment based on mathematical model according to claim 1, characterized in that: The neural network deep learning technology trains the mathematical model by preprocessing the raw data of the fixed section acquisition unit, selecting a predetermined neural network structure to build a neural network model, setting a loss function and an optimizer, and training the neural network model by defining the loss function and evaluation indicators, and regularly evaluating the performance through adjustable hyperparameters to form a predetermined positioning model.

7. A method for positioning computer equipment based on mathematical model according to claim 6, characterized in that: The neural network deep learning technology uses real-time variables to perform data enhancement on images in a database when training a mathematical model, and uses a regularization strategy to reduce redundant parameters.

8. The method for positioning computer equipment based on mathematical model according to claim 1, characterized in that: A real-time positioning algorithm is introduced into the matching unit.