Navigation system stability evaluation method under simulated extreme condition
Through detailed division and pre -processing of the historical and real -time data of the navigation system, combine the LSTM model and path factors affect the sequence, and correct the deviation of the deviation through the accuracy difference model. , Improve the accuracy and reliability of assessment.
Patent Information
- Application Number
- CN202411809849.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-10
- Publication Date
- 2025-01-24
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing navigation system stability evaluation methods cannot fully simulate and cover all extreme situations under extreme conditions, resulting in a large gap between the evaluation results and actual performance, and there are still errors in accuracy.
By obtaining the historical navigation data and real -time navigation data of the navigation system, it is divided into recent data and historical data, and the navigation path data is divided into transition paths and general paths according to the path medium. The LSTM model is used for real -time data prediction, and comparison with historical data, the construction path factors affect the sequence to predict, and finally correct the deviation through the accuracy difference model, re -evaluate the stability of the navigation system.
It improves the accuracy and reliability of the stability assessment of the navigation system under extreme conditions, reduces the gap between the assessment results and actual performance, and enhances the comprehensiveness and dynamic adaptability of the assessment.
Smart Images
Figure CN119354237A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of navigation technology, and in particular to a method for evaluating the stability of a navigation system under simulated extreme conditions. Background Art
[0002] The main direction of navigation system stability assessment is focused on accuracy, involving multiple factors such as environmental adaptability, signal transmission path, multipath effect, sensor accuracy, data fusion algorithm and system response time. The advantage of existing assessment methods is that they can efficiently evaluate system performance and provide repeatable and comparable assessment results through standardized tests and quantitative indicators. However, these methods have significant defects under extreme conditions because they cannot fully simulate and cover all extreme situations, resulting in a large gap between the assessment results and the actual performance. In addition, the existing navigation system stability assessment methods still have certain errors in accuracy and need further improvement and optimization. Summary of the invention
[0003] In order to overcome the shortcoming of large errors in navigation system stability assessment under extreme conditions, the present invention provides a navigation system stability assessment method under simulated extreme conditions.
[0004] The technical implementation scheme of the present invention is: a method for evaluating the stability of a navigation system under simulated extreme conditions, comprising the following steps:
[0005] S1: Obtain historical navigation data of the navigation system, and divide it into recent data and historical data according to time sequence; at the same time, obtain navigation path data, and divide it into transition path and ordinary path according to path medium;
[0006] S2: acquiring real-time navigation system data, extracting navigation data of the same type as the real-time data from historical navigation data, comparing the real-time navigation data with the historical navigation data, obtaining navigation data of similar types, and evaluating the stability of the navigation system on the data;
[0007] S3: Obtain the navigation system stability assessment result and construct the path factor influence sequence; predict the real-time navigation process according to the path factor influence sequence, and the prediction range includes direct direction prediction and indirect direction prediction; compare the prediction result with the navigation result of the real-time navigation system to further evaluate the stability of the navigation system;
[0008] S4: Construct an accuracy difference model, use the model to correct the stability deviation of the navigation system, re-evaluate the stability of the navigation system, and finally obtain the navigation system stability evaluation result.
[0009] Preferably, the historical navigation data of the navigation system is obtained and divided into recent data and historical data in chronological order; at the same time, the navigation path data is obtained and divided into transition path and ordinary path according to the path medium, including: the historical navigation data and navigation path data include data cleaning and standardization processing; the recent data and historical data include, nearby data including but not limited to data t moments before real-time navigation system navigation, data t1 hours before or data t2 working days before, historical data including but not limited to historical error data and historical correct data; the transition path includes stability path data in the data before and after stability of the navigation system, and the ordinary data includes other data except the transition path.
[0010] Preferably, the real-time navigation system data is obtained, navigation data of the same type as the real-time data is extracted from the historical navigation data, and the real-time navigation data is compared with the historical navigation data to obtain navigation data of similar types, including: firstly, the real-time navigation system data is predicted by using the LSTM model to obtain the real-time navigation system related data; secondly, data having the same path as the real-time navigation system data is extracted from the recent data, and if the data does not exist, data having a similar path as the real-time navigation system data is extracted from the recent data; finally, data is extracted from the historical data according to the recent data extraction method; the same path data extraction is calculated according to the following formula,
[0011]
[0012] x=e k+z
[0013] , where tanh represents normalization, k and z are the combination of extreme conditions in the two paths, k is the evaluation value of navigation system stability under extreme conditions predicted by real-time navigation data, and z is the evaluation value of navigation system stability under extreme conditions extracted from recent data. First, the z value that is time-related to k is verified, and then the z value that is not time-related to k is verified. [-1-0] means irrelevant, and [0-1] means relevant; similar path data extraction is calculated according to the following formula,
[0014] x=e k+z
[0015] , where k1 and z1 are the combination of extreme conditions in the two paths at different times, respectively. k1 and z1 are also the evaluation values of the stability of the navigation system under extreme conditions. If k1 does not exist, that is, it is not predicted in the real-time navigation data, a new value is assigned to k1 using the genetic model. Similarly, z1 is the extreme condition extracted from the most recent data. The subsequent verification rules are the same as z. The difference is that if z1 does not exist, a new value is also assigned to z1 according to the genetic model. x represents the correlation of the stability of the navigation system under the combination of extreme conditions.
[0016] Preferably, the navigation system stability assessment of these data includes: comparing the stability of the same path data with the real-time navigation system data, and using the comparison result as a stability misjudgment data set; comparing the stability of similar path data with the real-time navigation system data, and using the comparison result as a stability error data set; fusing the stability misjudgment data set with the stability error data set as a fused data set, and using the fused data set as a navigation system stability assessment result.
[0017] Preferably, the navigation system stability assessment result is obtained, and a path factor influence sequence is constructed; the real-time navigation process is predicted according to the path factor influence sequence, and the prediction range includes direct direction prediction and indirect direction prediction, including: according to the navigation system stability assessment result, the path in the navigation process of the navigation system is weighted, and a path factor influence sequence is constructed according to the marking result; the real-time navigation process is predicted using the path factor influence sequence, and the Kalman model prediction result is used as the direct direction prediction, and the Kalman model prediction result and the path factor influence sequence are used as the indirect direction prediction; wherein the weight marking process is to first obtain the path data, then obtain the transition path and the ordinary path, and finally adjust the transition path and the ordinary path using the navigation system stability assessment result.
[0018] Preferably, the prediction result is compared with the navigation result of the real-time navigation system to further evaluate the stability of the navigation system, including: the prediction result includes a direct direction prediction and an indirect direction prediction, the direct direction prediction includes the direction of the navigation system at the next moment, and the indirect direction includes the deviation direction of the navigation system, the prediction result and the navigation result of the real-time navigation system are fused into a new data set and defined as a second fused data set, the fused data set is compared with the second fused data set, the stability of the navigation system is further evaluated based on the comparison result, and the stability of the navigation system is re-evaluated.
[0019] Preferably, the re-evaluation of the stability of the navigation system includes: obtaining a training set of a fused data set and a second fused data set, constructing an SVM classification model, inputting the training set data into the SVM classification model, and obtaining a trained SVM classification model; obtaining a stability evaluation result of the navigation system using the SVM classification model; the specific process is: if the similarity between the results of the fused data set and the second fused data set is less than a first threshold, then readjusting the fused data set; if the similarity between the results of the fused data set and the second fused data set is greater than the first threshold and less than the second threshold, then readjusting the misjudgment data set or the error data set in the fused data set; if the similarity between the results of the fused data set and the second fused data set is greater than the second threshold, then directly performing the final evaluation; repeating this process until the similarity between the results of the fused data set and the second fused data set is greater than the second threshold.
[0020] Preferably, the construction of the accuracy difference model, using the model to correct the navigation system stability deviation, re-evaluating the stability of the navigation system, and finally obtaining the navigation system stability evaluation result, includes: obtaining direct direction prediction results and indirect direction prediction results, extracting a training set of range and duration data of the direct direction prediction results and the indirect direction prediction results, constructing the accuracy difference model, inputting the training set data into the accuracy difference model, obtaining the trained accuracy difference model, and using the accuracy difference model to correct the navigation system stability deviation; and then using the range data and duration data to evaluate the stability of the navigation system, thereby obtaining the final navigation system stability evaluation result.
[0021] Preferably, the evaluating the stability of the navigation system using the range data and the duration data includes: performing a first evaluation using an accuracy stability calculation formula, and then performing a second evaluation using a navigation system stability evaluation formula; the accuracy stability calculation formula is as follows,
[0022]
[0023] , where W represents the stability of the accumulated accuracy of each path in the normalized navigation process, q=1 represents the first path, r represents the last path, n represents the direct direction accuracy range, T1 represents the duration of the direct direction accuracy range, n1 represents the indirect 1 direction accuracy range, T2 represents the duration of the indirect 1 direction accuracy range, n2 represents the indirect 2 direction accuracy range, and T3 represents the duration of the indirect 2 direction accuracy range; similarly, the formula is used to calculate the minimum value of the indirect 1 direction and the indirect 1 direction and the direct direction in terms of stability, and then the stable range of accuracy stability is obtained; the indirect 1 direction and the indirect 2 directions are the maximum values in the deviation direction.
[0024] Preferably, the secondary evaluation using the navigation system stability evaluation formula includes: the navigation stability evaluation calculation formula is as follows,
[0025]
[0026] , where k1 and k2 are weight coefficients, x1 and x2 represent the correlation between the stability of the navigation system under extreme conditions under the same and similar paths, W1 and W2 are the maximum and minimum values of the accuracy stability, v1 and v2 are the start and end of navigation, and L is stability. The formula calculates the stability under extreme conditions from the beginning to the end of navigation.
[0027] Beneficial effects: First, by carefully dividing and preprocessing historical data and real-time data, the quality and representativeness of the basic data for evaluation are ensured, laying a solid foundation for subsequent evaluation. Secondly, the LSTM model is used to predict real-time data and compare it with historical data, which not only improves the accuracy of data matching, but also enhances the comprehensiveness and accuracy of the evaluation. Furthermore, the construction of the path factor influence sequence and the direct and indirect direction prediction significantly enhances the accurate evaluation of the importance and influencing factors of the navigation path, and improves the accuracy and reliability of the prediction. In addition, by comparing the prediction results with the actual navigation results and dynamically adjusting the evaluation results, the dynamic adaptability of the evaluation process is ensured, and the scientificity and rationality of the evaluation are further improved. Next, the SVM classification model is used to evaluate the stability of the navigation system. By setting different thresholds to adjust the fusion data set hierarchically, the precision and comprehensiveness of the evaluation results are ensured, and the reliability and stability of the evaluation results are effectively guaranteed. Finally, the accuracy difference model is constructed to correct the stability deviation of the navigation system, which reduces the difference between the prediction results and the actual results, improves the accuracy of the evaluation, and ensures the comprehensiveness and reliability of the evaluation results by comprehensively evaluating the performance of the navigation system. This series of steps not only improves the accuracy of the stability assessment of the navigation system under extreme conditions, but also provides strong support for the optimization and improvement of the navigation system, ultimately improving the adaptability and reliability of the navigation system and ensuring navigation performance in various complex environments. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 A flowchart of a method for evaluating the stability of a navigation system under simulated extreme conditions according to the present invention;
[0029] Figure 2 A schematic diagram of a process for re-evaluating the stability of a navigation system according to the present invention. DETAILED DESCRIPTION
[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0031] A method for evaluating the stability of a navigation system under simulated extreme conditions, such as Figure 1 As shown, the following steps are included:
[0032] S1: Obtain historical navigation data of the navigation system, and divide it into recent data and historical data according to time sequence; at the same time, obtain navigation path data, and divide it into transition path and ordinary path according to path medium;
[0033] S2: acquiring real-time navigation system data, extracting navigation data of the same type as the real-time data from historical navigation data, comparing the real-time navigation data with the historical navigation data, obtaining navigation data of similar types, and evaluating the stability of the navigation system on the data;
[0034] S3: Obtain the navigation system stability assessment result and construct the path factor influence sequence; predict the real-time navigation process according to the path factor influence sequence, and the prediction range includes direct direction prediction and indirect direction prediction; compare the prediction result with the navigation result of the real-time navigation system to further evaluate the stability of the navigation system;
[0035] S4: Construct an accuracy difference model, use the model to correct the stability deviation of the navigation system, re-evaluate the stability of the navigation system, and finally obtain the navigation system stability evaluation result.
[0036] Acquire historical navigation data of the navigation system, and divide it into recent data and historical data in chronological order; at the same time, acquire navigation path data, and divide it into transition path and ordinary path according to path medium, including: the historical navigation data and navigation path data include data cleaning and standardization processing; the recent data and historical data include, nearby data including but not limited to data t moments before navigation of the real-time navigation system, data t1 hours before or data t2 working days before, and historical data including but not limited to historical error data and historical correct data; the transition path includes stability path data in the data before and after stability of the navigation system, and the ordinary data includes other data except the transition path.
[0037] Further explanation is that the stability path data is obtained according to the SVM model; the recent data (such as the data of the previous 10 minutes, the previous hour or the previous working day of real-time navigation) ensures the accurate reflection of the current state, and the historical data (including historical error data and historical correct data) provides a comprehensive performance reference, and the timeliness and representativeness of the data are guaranteed; the key transition paths (such as major intersections and tunnel entrances in the city) are identified through the SVM model, which helps to more accurately identify the stability of the navigation system on these paths, while the ordinary paths include other paths except the transition paths, which improves the accuracy of the evaluation. For example, the main intersections in the city are identified as stability path data through the SVM model. These paths are prone to multipath effects and signal interference during the navigation process, which provides strong support for system optimization and improvement.
[0038] Acquire real-time navigation system data, extract navigation data of the same type as the real-time data from historical navigation data, compare the real-time navigation data with these historical navigation data, and obtain navigation data of similar types, including: first, use the LSTM model to predict the real-time navigation system data to obtain real-time navigation system related data, and then extract data with the same path as the real-time navigation system data from the latest data. If the data does not exist, extract the data with a similar path as the real-time navigation system data from the latest data; finally, extract from the historical data according to the latest data extraction method; the same path data extraction is calculated according to the following formula,
[0039]
[0040] x=e k+z
[0041] , where tanh represents normalization, k and z are the combination of extreme conditions in the two paths, k is the evaluation value of the stability of the navigation system under extreme conditions predicted by real-time navigation data, k1 and z1 are also the evaluation values of the stability of the navigation system under extreme conditions, z is the evaluation value of the stability of the navigation system under extreme conditions extracted from the most recent data, first verify the z value that is time-related to k, and then verify the z value that is not time-related to k, [-1-0] means irrelevant, and [0-1] means relevant; similar path data extraction is calculated according to the following formula,
[0042] x=e k1+z1
[0043] , where k1 and z1 are the combination of extreme conditions in the two paths at different times. If k1 does not exist, that is, it is not predicted in the real-time navigation data, the genetic model is used to assign a new value to k1. Similarly, z1 is the extreme condition extracted from the most recent data. The subsequent verification rules are the same as z. The difference is that if z1 does not exist, a new value is also assigned to z1 according to the genetic model. x represents the correlation of the stability of the navigation system under the combination of extreme conditions.
[0044] Further explanation is that the real-time navigation system data is first predicted using the LSTM model to ensure the accuracy and reliability of the prediction. By extracting data of the same or similar paths from recent data, the data matching is refined and comprehensive. The scientific calculation of path similarity uses formulas to verify the combination of extreme conditions in the path to ensure the scientificity and rationality of path similarity. Under extreme conditions, new values are assigned using genetic models to ensure data processing capabilities and comprehensively improve the stability and reliability of the navigation system. For example, the LSTM model is used to predict the current traffic congestion of vehicles at major intersections, extract data of the same path from the data of the last 10 minutes, and extract data of similar paths if it does not exist, and calculate path similarity through formulas, ensuring that the stability evaluation of the navigation system under different paths and extreme conditions is more accurate and reliable.
[0045] The navigation system stability is evaluated on these data, including: comparing the stability of the same path data with the real-time navigation system data, and using the comparison result as the stability misjudgment data set; comparing the stability of similar path data with the real-time navigation system data, and using the comparison result as the stability error data set; fusing the stability misjudgment data set with the stability error data set as a fused data set, and using the fused data set as the navigation system stability evaluation result.
[0046] A further explanation is that by comparing the same path data with the real-time navigation system data, the stability of the navigation system on the same path is accurately evaluated, thereby improving the accuracy of the evaluation. The comprehensiveness of the evaluation is increased by comparing similar path data to ensure a comprehensive evaluation of the stability of similar paths. The reliability of the fused data set is improved by fusing the stability misjudgment data set with the error data set to improve the reliability and comprehensiveness of the evaluation results. Misjudgments and errors are identified, and misjudgment and error data sets are generated to provide a basis for system optimization. For example, by comparing the navigation data of the current vehicle at the main intersection A with historical data, a stability misjudgment data set is generated; by comparing the navigation data close to the main intersection A with historical data, a stability error data set is generated; and finally, the two data sets are fused as the evaluation result to ensure that the stability evaluation of the navigation system on different paths is more accurate and reliable.
[0047] The method comprises obtaining a navigation system stability assessment result, constructing a path factor influence sequence; predicting a real-time navigation process according to the path factor influence sequence, wherein the prediction range includes direct direction prediction and indirect direction prediction, including: weighting the path in the navigation process of the navigation system according to the navigation system stability assessment result, and constructing a path factor influence sequence according to the marking result; predicting the real-time navigation process using the path factor influence sequence, using the Kalman model prediction result as the direct direction prediction, and using the Kalman model prediction result and the path factor influence sequence as the indirect direction prediction; wherein the weight marking process is to first obtain path data, then obtain a transition path and an ordinary path, and finally adjust the transition path and the ordinary path using the navigation system stability assessment result.
[0048] Further explanation is that by weighting the path and constructing a path factor influence sequence, we can ensure accurate evaluation of the importance and influencing factors of the path and improve prediction accuracy. We use the path factor influence sequence and the Kalman model to make comprehensive predictions, dynamically adjust the path weights, optimize navigation decisions, and improve the stability and reliability of the navigation system. For example, by evaluating the stability of the main intersection A as 0.8 and the stability of the ordinary road section B as 0.9, we weight the path and construct a path factor influence sequence; use the Kalman model to predict the direct direction of the vehicle at the next intersection, and combine the path factor influence sequence to predict the indirect direction of the vehicle when approaching the main intersection A, ensuring that the navigation system's predictions on different paths are more accurate and reliable.
[0049] The prediction results are compared with the navigation results of the real-time navigation system to further evaluate the stability of the navigation system, including: the prediction results include direct direction prediction and indirect direction prediction, the direct direction prediction includes the direction of the navigation system at the next moment, and the indirect direction includes the deviation direction of the navigation system; the prediction results are fused with the navigation results of the real-time navigation system into a new data set and defined as a second fused data set; the fused data set is compared with the second fused data set; the stability of the navigation system is further evaluated based on the comparison results, and the stability of the navigation system is re-evaluated.
[0050] Further explanation is that the accuracy and reliability of the evaluation are ensured by comparing the prediction results (direct direction prediction and indirect direction prediction) with the navigation results of the real-time navigation system. The navigation system performance is comprehensively evaluated, including the direction and deviation direction at the next moment, to comprehensively evaluate the system performance. A new data set (the second fused data set) is generated to ensure the integrity and consistency of the data. The evaluation results are dynamically adjusted, and the stability of the navigation system is further evaluated by comparing the fused data set with the second fused data set to ensure the dynamic and adaptability of the evaluation results. For example, by predicting that the vehicle will turn right at the next intersection and choose to detour other paths when approaching the main intersection A, and comparing with the actual navigation results, a second fused data set is generated to confirm that the navigation system has high stability and ensure system performance and reliability.
[0051] Re-evaluating the stability of the navigation system, including: obtaining a training set of a fused data set and a second fused data set, constructing an SVM classification model, inputting the training set data into the SVM classification model, and obtaining a trained SVM classification model; obtaining a stability evaluation result of the navigation system with the SVM classification model; the specific process is: if the similarity between the results of the fused data set and the second fused data set is less than a first threshold, then readjusting the fused data set; if the similarity between the results of the fused data set and the second fused data set is greater than the first threshold and less than the second threshold, then readjusting the misjudgment data set or the error data set in the fused data set; if the similarity between the results of the fused data set and the second fused data set is greater than the second threshold, then directly performing the final evaluation; repeating this process until the similarity between the results of the fused data set and the second fused data set is greater than the second threshold.
[0052] Further explanation is that by constructing an SVM classification model, the accuracy and reliability of the evaluation results are ensured. According to the similarity between the fused data set and the second fused data set, the fused data set is dynamically adjusted to ensure the dynamic and adaptability of the evaluation results. By setting different thresholds, the fused data set is adjusted hierarchically to ensure the precision and comprehensiveness of the evaluation results. The adjustment process is repeated until the similarity is greater than the second threshold to ensure the reliability and stability of the final evaluation results. For example, suppose that the stability of a navigation system in an urban environment needs to be re-evaluated. First, a data training set is obtained, including a fused data set (fusion data of the predicted results and the actual navigation results) and a second fused data set (comparison data between the navigation results of the real-time navigation system and the predicted results). An SVM classification model is constructed, and the training set data is input into the model to obtain a trained SVM classification model. The SVM classification model is used for initial evaluation to calculate the similarity between the fused data set and the second fused data set. Dynamically adjust the fused data set according to the similarity: if the similarity is less than the first threshold (e.g. 0.6), readjust the fused data set; if the similarity is greater than the first threshold and less than the second threshold (e.g. 0.6 to 0.8), adjust the misjudgment data set or the error data set; if the similarity is greater than the second threshold (e.g. 0.8), directly perform the final evaluation. In the specific example, by predicting that the vehicle will turn right at the next intersection and choose to take other detours when approaching the main intersection A, and comparing it with the actual navigation results, a second fused data set is generated, and finally the stability score of the navigation system is confirmed to be 0.85, ensuring the performance and reliability of the navigation system.
[0053] A precision difference model is constructed, and the model is used to correct the navigation system stability deviation, and the stability of the navigation system is re-evaluated to finally obtain the navigation system stability evaluation result, including: obtaining direct direction prediction results and indirect direction prediction results, extracting the training set of range and duration data of the direct direction prediction results and the indirect direction prediction results, constructing the precision difference model, inputting the training set data into the precision difference model, obtaining the trained precision difference model, and correcting the navigation system stability deviation using the precision difference model; and then using the range data and duration data to evaluate the stability of the navigation system, thereby obtaining the final navigation system stability evaluation result.
[0054] Further explanation is that by constructing the precision difference model and using the training set data for training, the accuracy and reliability of the evaluation results are ensured. The precision difference model is used to correct the stability deviation of the navigation system, reduce the difference between the predicted results and the actual results, and improve the accuracy of the evaluation. The range and duration data of the direct direction prediction results and the indirect direction prediction results are extracted to comprehensively evaluate the performance of the navigation system and ensure the comprehensiveness and reliability of the evaluation results. The dynamic and adaptability of the final evaluation results are ensured by repeatedly using the precision difference model for correction and evaluation. For example, suppose that the stability of a navigation system in an urban environment needs to be evaluated. Direct direction predictions (such as predicting that the vehicle will turn right at the next intersection for 10 seconds) and indirect direction predictions (such as predicting that the vehicle will choose to detour to other paths when approaching the main intersection A for 30 seconds) are obtained, and the range and duration data are extracted to build and train the precision difference model. The deviation is corrected by the model and the stability of the navigation system is re-evaluated. The final evaluation result shows that the stability score of the navigation system is 0.9, indicating that the navigation system performs well in an urban environment.
[0055] The range data and duration data are used to evaluate the stability of the navigation system, including: using the accuracy stability calculation formula for a first evaluation, and then using the navigation system stability evaluation formula for a second evaluation; the accuracy stability calculation formula is as follows,
[0056]
[0057] , where W represents the stability of the accumulated accuracy of each path in the normalized navigation process, q=1 represents the first path, r represents the last path, n represents the direct direction accuracy range, T1 represents the duration of the direct direction accuracy range, n1 represents the indirect 1 direction accuracy range, T2 represents the duration of the indirect 1 direction accuracy range, n2 represents the indirect 2 direction accuracy range, and T3 represents the duration of the indirect 2 direction accuracy range; similarly, the formula is used to calculate the minimum value of the indirect 1 direction and the indirect 1 direction and the direct direction in terms of stability, and then the stable range of accuracy stability is obtained; the indirect 1 direction and the indirect 2 directions are the maximum values in the deviation direction.
[0058] Further explanation is that the accuracy of the navigation system on each path is quantitatively evaluated through the accuracy stability calculation formula to ensure the accuracy and reliability of the evaluation results. The accuracy range and duration data of the direct direction and indirect direction are used to comprehensively evaluate the performance of the navigation system to ensure the comprehensiveness and reliability of the evaluation results. Through normalization processing, the stability of the accumulated accuracy of each path is standardized for easy comparison and analysis. By calculating the minimum value of the indirect direction and the direct direction in terms of stability, the stability range of the accuracy stability is determined to ensure the scientificity and rationality of the evaluation results. For example, assuming that the stability of a navigation system in an urban environment is evaluated, the direct direction accuracy range (2 meters, 10 seconds), the indirect 1 direction accuracy range (3 meters, 30 seconds) and the indirect 2 direction accuracy range (4 meters, 20 seconds) are obtained. The accuracy stability calculation formula is used for the first evaluation to calculate the normalized stability value. Then the navigation system stability evaluation formula is used for the second evaluation, and the stability range is finally determined to be 0.727 to 0.9, ensuring the scientificity and rationality of the evaluation results.
[0059] The navigation system stability evaluation formula is used for secondary evaluation, including: The navigation stability evaluation calculation formula is as follows,
[0060]
[0061] , where k1 and k2 are weight coefficients, x1 and x2 represent the correlation between the stability of the navigation system under extreme conditions under the same and similar paths, W1 and W2 are the maximum and minimum values of the accuracy stability, v1 and v2 are the start and end of navigation, and L is stability. The formula calculates the stability under extreme conditions from the beginning to the end of navigation.
[0062] The above is a detailed introduction to the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
Claims
1. A method for evaluating the stability of a navigation system under simulated extreme conditions, characterized in that: The following steps are involved: S1: Obtain historical navigation data of the navigation system, and divide it into recent data and historical data according to time sequence; at the same time, obtain navigation path data, and divide it into transition path and ordinary path according to path medium; S2: acquiring real-time navigation system data, extracting navigation data of the same type as the real-time data from historical navigation data, comparing the real-time navigation data with the historical navigation data, obtaining navigation data of similar types, and evaluating the stability of the navigation system on the data; S3: Obtain navigation system stability assessment results and construct path factor influence sequence; Predict the real-time navigation process based on the path factor influence sequence, including direct direction prediction and indirect direction prediction; compare the prediction results with the navigation results of the real-time navigation system to further evaluate the stability of the navigation system; S4: Construct an accuracy difference model, use the model to correct the stability deviation of the navigation system, re-evaluate the stability of the navigation system, and finally obtain the navigation system stability evaluation result.
2. A navigation system stability assessment method under simulated extreme conditions according to claim 1, characterized in that: The method of obtaining the historical navigation data of the navigation system and dividing it into the latest data and historical data in chronological order; at the same time, obtaining the navigation path data and dividing it into the transition path and the ordinary path according to the path medium, includes: the historical navigation data and the navigation path data include data cleaning and standardization processing; the recent data and the historical data include, the nearby data include but are not limited to the data of the real-time navigation system before t moments, the data of the previous t1 hours or the previous t2 working days, and the historical data include but are not limited to the historical error data and the historical correct data; the transition path includes the stability path data in the data before and after the stability of the navigation system, and the ordinary data includes other data except the transition path.
3. A navigation system stability assessment method under simulated extreme conditions according to claim 1, characterized in that: The real-time navigation system data is obtained, navigation data with the same type as the real-time data is extracted from the historical navigation data, and the real-time navigation data is compared with the historical navigation data to obtain navigation data of similar types, including: firstly, the real-time navigation system data is predicted by using the LSTM model to obtain the real-time navigation system related data; secondly, data with the same path as the real-time navigation system data is extracted from the recent data, and if the data does not exist, data with a similar path as the real-time navigation system data is extracted from the recent data; finally, data is extracted from the historical data according to the recent data extraction method; the same path data extraction is calculated according to the following formula, x=e k+z , Among them, tanh represents normalization, k and z are the combination of extreme conditions in the two paths, k is the evaluation value of the stability of the navigation system under extreme conditions predicted by real-time navigation data, and z is the evaluation value of the stability of the navigation system under extreme conditions extracted from the most recent data. First, the z value that is time-related to k is verified, and then the z value that is not time-related to k is verified. [-1-0] means irrelevant, and [0-1] means relevant; similar path data extraction is calculated according to the following formula, x=e k1+z1 , Among them, k1 and z1 are the combination of extreme conditions in two paths at different times, and k1 and z1 are also the evaluation values of the stability of the navigation system under extreme conditions. If k1 does not exist, that is, it is not predicted in the real-time navigation data, the genetic model is used to assign a new value to k1. Similarly, z1 is the extreme condition extracted from the most recent data. The subsequent verification rules are the same as z. The difference is that if z1 does not exist, a new value is also assigned to z1 according to the genetic model. x represents the correlation of the stability of the navigation system under the combination of extreme conditions.
4. A navigation system stability assessment method under simulated extreme conditions according to claim 1, characterized in that: The navigation system stability assessment of these data includes: comparing the stability of the same path data with the real-time navigation system data, and using the comparison result as a stability misjudgment data set; comparing the stability of similar path data with the real-time navigation system data, and using the comparison result as a stability error data set; fusing the stability misjudgment data set with the stability error data set as a fused data set, and using the fused data set as a navigation system stability assessment result.
5. A navigation system stability assessment method under simulated extreme conditions according to claim 1, characterized in that: The navigation system stability assessment result is obtained, and a path factor influence sequence is constructed; The real-time navigation process is predicted according to the path factor influence sequence, and the prediction scope includes direct direction prediction and indirect direction prediction, including: according to the navigation system stability evaluation result, weight marking is performed on the path in the navigation process of the navigation system, and the path factor influence sequence is constructed according to the marking result; the real-time navigation process is predicted by using the path factor influence sequence, and the Kalman model prediction result is used as the direct direction prediction, and the Kalman model prediction result and the path factor influence sequence are used as the indirect direction prediction; wherein the weight marking process is to first obtain the path data, then obtain the transition path and the ordinary path, and finally adjust the transition path and the ordinary path by using the navigation system stability evaluation result.
6. A navigation system stability assessment method under simulated extreme conditions according to claim 1, characterized in that: The prediction result is compared with the navigation result of the real-time navigation system to further evaluate the stability of the navigation system, including: the prediction result includes a direct direction prediction and an indirect direction prediction, the direct direction prediction includes the direction of the navigation system at the next moment, and the indirect direction includes the deviation direction of the navigation system; the prediction result and the navigation result of the real-time navigation system are fused into a new data set and defined as a second fused data set; the fused data set is compared with the second fused data set, the stability of the navigation system is further evaluated according to the comparison result, and the stability of the navigation system is re-evaluated.
7. A navigation system stability assessment method under simulated extreme conditions according to claim 6, characterized in that: The re-evaluation of the stability of the navigation system includes: obtaining a training set of a fused data set and a second fused data set, constructing an SVM classification model, inputting the training set data into the SVM classification model, and obtaining a trained SVM classification model; obtaining a stability evaluation result of the navigation system with the SVM classification model; the specific process is: if the similarity between the fused data set and the second fused data set result is less than a first threshold, then readjusting the fused data set; if the similarity between the fused data set and the second fused data set result is greater than the first threshold and less than the second threshold, then readjusting the misjudgment data set or the error data set in the fused data set; if the similarity between the fused data set and the second fused data set result is greater than the second threshold, then directly performing the final evaluation; repeating this process until the similarity between the fused data set and the second fused data set result is greater than the second threshold.
8. The method for evaluating the stability of a navigation system under simulated extreme conditions according to claim 1, characterized in that: The method of constructing an accuracy difference model, using the model to correct the navigation system stability deviation, re-evaluating the stability of the navigation system, and finally obtaining a navigation system stability evaluation result includes: obtaining a direct direction prediction result and an indirect direction prediction result, extracting a training set of range and duration data of the direct direction prediction result and the indirect direction prediction result, constructing an accuracy difference model, inputting the training set data into the accuracy difference model, obtaining a trained accuracy difference model, and correcting the navigation system stability deviation using the accuracy difference model; and then evaluating the stability of the navigation system using the range data and the duration data, thereby obtaining a final navigation system stability evaluation result.
9. A navigation system stability assessment method under simulated extreme conditions according to claim 8, characterized in that: The use of range data and duration data to evaluate the stability of the navigation system includes: using an accuracy stability calculation formula to perform a first evaluation, and then using a navigation system stability evaluation formula to perform a second evaluation; the accuracy stability calculation formula is as follows, ， Among them, W represents the stability of the accumulated accuracy of each path in the normalized navigation process, q=1 represents the first path, r represents the last path, n represents the direct direction accuracy range, T1 represents the duration of the direct direction accuracy range, n1 represents the indirect 1 direction accuracy range, T2 represents the duration of the indirect 1 direction accuracy range, n2 represents the indirect 2 direction accuracy range, and T3 represents the duration of the indirect 2 direction accuracy range; similarly, the formula is used to calculate the minimum value of the indirect 1 direction, the indirect 1 direction and the direct direction in terms of stability, and then the stable range of accuracy stability is obtained; the indirect 1 direction and the indirect 2 directions are the maximum values in the deviation direction.
10. A navigation system stability assessment method under simulated extreme conditions according to claim 9, characterized in that: The secondary evaluation using the navigation system stability evaluation formula includes: the navigation stability evaluation calculation formula is as follows: ， Among them, k1 and k2 are weight coefficients, x1 and x2 represent the correlation between the stability of the navigation system under the combination of extreme conditions on the same path and similar paths, W1 and W2 are the maximum and minimum values of the accuracy stability, v1 and v2 are the start and end of navigation, L is stability, and the formula calculates the stability under extreme conditions changes from the beginning to the end of navigation.