A method for monitoring the flight status of unmanned aerial vehicles for air-ground integration

Abstract

The present invention belongs to the field of data processing technology, and specifically relates to a method for monitoring the flight status of an unmanned aerial vehicle for air-ground integration, the method comprising: performing curve fitting on each pitch angle data point in the attitude data of the unmanned aerial vehicle, determining the irregular volatility of the pitch angle data point according to the fitting error, and determining the noise possibility according to the difference between the pitch angle data point and the irregular volatility of the yaw angle data point and the roll angle data point at the same time; performing iterative SG filtering on the pitch angle sequence, determining the preferred degree of each iteration according to the noise possibility and the filtering amount of each iteration of the pitch angle data point, setting the iteration stop condition according to the preferred degree, and obtaining the final filtering result; and monitoring the flight status of the unmanned aerial vehicle according to the final filtering results of the pitch angle sequence, the yaw angle sequence, and the roll angle sequence. The present invention can significantly improve the filtering effect of the attitude data and improve the accuracy of the attitude monitoring of the unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology. More specifically, the present invention relates to a flight status monitoring method for an air-ground integrated unmanned aerial vehicle. Background Art

[0002] In recent years, drone technology has made significant progress. Drones are not only used in aerial photography, logistics, agriculture and other fields, but also play an important role in environmental monitoring and ecological protection. With the continuous maturity and popularization of drone technology, the demand for real-time monitoring of drone flight status is becoming increasingly urgent. Among them, drone attitude data (pitch angle, yaw angle, roll angle) directly reflects the flight attitude and stability of the drone. By monitoring these data in real time, abnormal attitudes of drones during flight, such as excessive tilt and yaw, can be discovered in time, so that measures can be taken immediately to correct them and avoid flight accidents.

[0003] In order to accurately monitor the flight status of the UAV, it is necessary to filter and denoise the collected attitude data. The paper (CHEN, JA simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky–Golay filter[J]. Remote Sensing of Environment, 2004) proposed a method for iterative SG filtering of time series data sequences. The algorithm provides a more accurate filtering effect by observing the fitting error index during the iteration process.

[0004] Since there are inevitably some normal attitude changes in the operation of the UAV, which cause drastic changes in its attitude data, such changes are similar to those caused by noise. The termination condition of the existing iterative SG filtering method is determined by the change of the fitting error index of three consecutive iterations. This method may lead to local optimality, that is, filtering or under-filtering problems may appear in the filtering results, thereby affecting the accuracy of UAV flight status monitoring. Summary of the invention

[0005] In order to solve the technical problem that the iterative SG filtering of the UAV attitude data may cause over-filtering or under-filtering, thus affecting the accuracy of the UAV flight status monitoring, the present invention provides a UAV flight status monitoring method for air-ground integration, comprising:

[0006] The attitude data of the UAV is collected, including a pitch angle sequence, a yaw angle sequence, and a roll angle sequence; curve fitting is performed on data points within a local range of each pitch angle data point in the pitch angle sequence, and the irregular volatility of each pitch angle data point is determined according to the fitting error; the noise possibility of the pitch angle data point is determined according to the difference between the irregular volatility of the pitch angle data point and the irregular volatility of the yaw angle data point and the roll angle data point at the same time; iterative SG filtering is performed on the pitch angle sequence, and the preferred degree of each iteration is determined according to the correlation between the noise possibility of each pitch angle data point in the pitch angle sequence and the filtering amount of each pitch angle data point in each iteration process; in response to the preferred degree being greater than a preset stop threshold, the iteration is stopped to obtain the final filtering result of the pitch angle sequence; the flight status of the UAV is monitored according to the final filtering results of the pitch angle sequence, the yaw angle sequence, and the roll angle sequence.

[0007] The effect is that: the present invention can effectively distinguish normal attitude changes from noise caused by equipment, environment and other factors by curve fitting the pitch angle sequence, determining the irregular fluctuation of the pitch angle data points based on the fitting error, and identifying potential noise in combination with the irregular fluctuation of the yaw angle data points and the roll angle data points at the same time. This identification mechanism helps to more accurately eliminate or correct invalid data and optimize the response of the flight control and monitoring system; the present invention uses an iterative SG filtering method to determine the preferred degree of iteration by calculating the relationship between the noise possibility and the filtering amount of each data point, and can intelligently optimize the filtering process according to the noise characteristics of the data, avoid excessive smoothing and cause data information loss, and can retain more effective information, thereby improving the accuracy of the final filtering result of the UAV attitude data; the pitch angle, yaw angle, and roll angle sequence after the final filtering can more accurately reflect the real flight state of the UAV, reduce the misjudgment caused by data noise, and thus improve the accuracy and reliability of flight state monitoring.

[0008] Preferably, the elements in the pitch angle sequence, the yaw angle sequence and the roll angle sequence are pre-processed data of the pitch angle, the yaw angle and the roll angle, and the pre-processed data are within the range.

[0009] The effect is that normalizing the angle data to the range of [0,1] can eliminate the scale differences of the original angle values, so that data of different dimensions (pitch angle, yaw angle, roll angle) can be processed on the same scale, avoiding deviations in subsequent processing due to excessively large or small values ​​of certain angle values, thereby improving the calculation efficiency and stability of the results.

[0010] Preferably, the irregular fluctuation of the pitch angle data points satisfies the expression:

[0011] ; In the formula, Indicates the pitch angle data point The irregular fluctuations of is the first Pitch angle data points; Indicates the pitch angle data point The number of data points in the local range; Indicates the pitch angle data point In local area data points; Indicates the pitch angle data point In local area The fitted value of the data points; Indicates the pitch angle data point In local area The fitting error of the data points; is the absolute value symbol.

[0012] The effect is that irregular volatility quantifies the average deviation of the pitch angle data points in the local range relative to their fitted values, which helps to analyze the fluctuation intensity and regularity of the pitch angle data series in different local ranges.

[0013] Preferably, the noise probability of the pitch angle data point satisfies the expression:

[0014] ; In the formula, Indicates the pitch angle data point The possibility of noise, Indicates the pitch angle data point Irregular volatility; , Respectively represent the data points with pitch angle Irregular fluctuations of yaw angle data points and roll angle data points at the same time; Represents an exponential function with a natural constant as its base.

[0015] The effect is that based on the difference between the pitch angle data points and the irregular fluctuations of the yaw angle data points and the roll angle data points at the same time, it is possible to effectively distinguish between normal attitude changes and noise caused by equipment, environment and other factors, which helps to filter the noise data more accurately later.

[0016] Preferably, the method for obtaining the filtering amount is: performing iterative SG filtering on the pitch angle sequence to obtain the first The result of the iteration is recorded as , convert the sequence Subtract the pitch angle sequence from the pitch angle sequence to obtain the first The amount of filtering for each iteration.

[0017] Preferably, the method of determining the degree of preference of each iteration according to the correlation between the noise possibility of each pitch angle data point in the pitch angle sequence and the filtering amount of each pitch angle data point in each iteration process includes: establishing a rectangular coordinate system with the noise possibility as the horizontal axis and the filtering amount as the vertical axis, and drawing a scatter plot according to the noise possibility and the filtering amount of each pitch angle data point in the pitch angle sequence; performing linear fitting on the points in the scatter plot to obtain a linear regression curve; and determining the degree of preference of this iteration according to the average distance from each point in the scatter plot to the linear regression curve, wherein the average distance reflects the correlation between the noise possibility and the filtering amount.

[0018] The effect is that the linear regression curve can reveal the quantitative relationship between the noise possibility and the filtering amount, making the filtering iteration more targeted and achieving the optimal filtering effect.

[0019] Preferably, the preference degree satisfies the expression: ;in, Indicates the first The optimization degree of the iteration, Indicates After iterations, the average distance from all points to the linear regression curve in the scatter plot established based on the noise probability of each pitch angle data point in the pitch angle sequence and the filtering amount is: is an exponential function with a natural constant as its base.

[0020] Preferably, the method further comprises: after filtering is completed, restoring the final filtering result of the pitch angle sequence to Within the range, restore the final filtering results of the yaw angle data sequence and the final filtering results of the roll angle sequence to within the range.

[0021] Preferably, the preprocessing process is: the preprocessing process for the pitch angle is: ,in is the pitch angle obtained, is the preprocessed pitch angle data, is a sine function; the preprocessing process of the yaw angle is: ,in is the collected yaw angle, is the preprocessed pitch angle data; the preprocessing process for the roll angle is: ,in is the collected roll angle, is the preprocessed roll angle data.

[0022] Preferably, the restoration process is: the restoration process of the pitch angle is: ,in is the pitch angle after filtering, is the final filtering result of the preprocessed pitch angle data, is the inverse sine function; the restoration process of the yaw angle is: ,in is the yaw angle after filtering, is the final filtering result of the preprocessed yaw angle data; the restoration process of the roll angle is: ,in is the roll angle after filtering, is the final filtering result of the preprocessed roll angle data.

[0023] The beneficial effects of the present invention are:

[0024] 1. The noise probability in the present invention can effectively distinguish normal attitude changes from noise caused by equipment, environment and other factors, which helps to more accurately eliminate or correct invalid data and optimize the response of flight control and monitoring systems;

[0025] 2. The present invention obtains the optimization degree of iteration by quantifying the relationship between the noise possibility and the filtering amount in each iteration process, and sets the iteration stop condition according to the optimization degree, which can avoid the iterative SG filtering from falling into the local optimum, avoid overfitting or underfitting, retain more effective information, and improve the accuracy of the final filtering result of the UAV attitude data;

[0026] 3. The pitch angle, yaw angle, and roll angle sequences finally filtered by the present invention can more accurately reflect the actual flight status of the UAV, reduce misjudgment caused by data noise, and improve the accuracy and reliability of UAV flight status monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below through the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0028] Figure 1 is a flow chart schematically showing a method for monitoring the flight status of an air-to-ground integrated UAV in the present invention;

[0029] Figure 2 is a diagram schematically showing the drone attitude data;

[0030] Figure 3 It is a flow chart schematically showing step S2 of a method for monitoring the flight status of an air-to-ground integrated UAV in the present invention. DETAILED DESCRIPTION

[0031] 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 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0032] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0033] The embodiment of the present invention discloses a method for monitoring the flight status of an air-ground integrated unmanned aerial vehicle, referring to Figure 1 , including steps S1 to S3:

[0034] S1. Collect the attitude data of the drone, including pitch angle sequence, yaw angle sequence and roll angle sequence.

[0035] Specifically, the fusion method of accelerometer and magnetometer is used to collect the UAV attitude data in real time, including the pitch angle, yaw angle and roll angle of the UAV. In this embodiment, the collection frequency is 2 Hz. In other embodiments, the implementer can set the collection frequency according to the actual implementation situation.

[0036] like Figure 2 As shown, the pitch angle of the drone describes the rotation angle of the object around the horizontal axis (x-axis), and the value range is ; The yaw angle describes the rotation angle of the object around the vertical axis (z axis), and its value range is , the roll angle describes the rotation angle of the object around the longitudinal axis (y axis), and its value range is Spend.

[0037] It should be noted that in order to facilitate the subsequent analysis of the attitude data, the collected UAV attitude data needs to be preprocessed, and the pitch angle, yaw angle, and roll angle are limited to within the range.

[0038] In one embodiment, the preprocessing process of the pitch angle is: ,in is the pitch angle obtained, is the preprocessed pitch angle data, is a sine function;

[0039] The preprocessing process of the yaw angle is: ,in is the collected yaw angle, is the preprocessed pitch angle data;

[0040] The preprocessing process of the roll angle is: ,in is the collected roll angle, is the preprocessed roll angle data;

[0041] The above processing can limit the pitch angle, yaw angle, and roll angle to The preprocessing results are within the range and different angle values ​​are satisfied, which facilitates the subsequent analysis and processing of the UAV attitude data.

[0042] In another embodiment, the preprocessing process of the pitch angle is: ,in is the pitch angle obtained, is the preprocessed pitch angle data;

[0043] The preprocessing process of the yaw angle is: ,in is the collected yaw angle, is the preprocessed pitch angle data;

[0044] The preprocessing process of the roll angle is: ,in is the collected roll angle, is the preprocessed roll angle data;

[0045] The above processing can limit the pitch angle, yaw angle, and roll angle to The preprocessing results are within the range and different angle values ​​are satisfied, which facilitates the subsequent analysis and processing of the UAV attitude data.

[0046] In other embodiments, the implementer may also use other methods to pre-process the pitch angle, yaw angle, and roll angle, and limit the pitch angle, yaw angle, and roll angle to within the range.

[0047] The pitch angle data pre-processed by the UAV is constructed into a pitch angle sequence in time sequence , the yaw angle data pre-processed by the UAV is arranged in time sequence to form a yaw angle sequence , the pre-processed roll data of the UAV is arranged in time sequence to form a roll angle sequence .

[0048] S2. Set the iteration stop condition of iterative SG filtering and perform iterative SG filtering on the attitude data of the UAV.

[0049] It should be noted that the traditional iterative SG filter determines whether to stop the iteration by the change of the fitting error index of three consecutive iterations, that is, when The iteration is terminated when , , Respectively represent , , The fitting error index at the iteration is terminated when the fitting error index reaches the minimum. This method may lead to a local optimum, that is, The effect of the first iteration is slightly better than that of the Second and There may be better filtering effects later, which leads to under-filtering, and this method cannot distinguish between normal data changes in drone attitude data and noise changes caused by the environment, resulting in filter waves. In response to the above problems, the present invention determines the noise possibility of attitude data points by analyzing the change characteristics of drone attitude data, obtains the preferred degree of the iterative SG filtering method in the iterative process according to the noise possibility, and sets the iterative stop condition of the iterative SG filtering method according to the preferred degree to avoid overfitting or underfitting. Specifically:

[0050] See the flowchart of step S2. Figure 3 , including steps S201 to S203:

[0051] S201 , performing curve fitting on data points within a local range of each pitch angle data point in the pitch angle sequence, and determining irregular fluctuations of each pitch angle data point according to a fitting error.

[0052] It should be noted that during the flight of the UAV, factors such as airflow will affect the operation of the accelerometer and magnetometer, resulting in noise in the collected UAV attitude data. Such noise manifests itself as irregular fluctuations in the local range of the attitude data. Therefore, the present invention uses the irregular volatility of the local range of the one-dimensional attitude data to calculate its noise possibility. The stronger the irregular volatility, the stronger the noise possibility.

[0053] Specifically, the method for obtaining irregular volatility includes:

[0054] The pitch angle sequence Middle The data points are recorded as , whose local range is based on the pitch angle data points centered data points, in this embodiment In other embodiments, the implementer can set .

[0055] Pitch angle data points The first data point in the local range is taken as the center, the time sequence of the data points in the local range is taken as the horizontal coordinate, and the data value is taken as the vertical coordinate to establish a rectangular coordinate system. The least squares method is used to calculate the pitch angle data points. The curve is fitted to the data points in the local area.

[0056] Determine the pitch angle data points based on the fitting results Irregular volatility:

[0057] ;

[0058] In the formula, Indicates the pitch angle data point Irregular volatility; Indicates the pitch angle data point The number of data points in the local range; Indicates the pitch angle data point In local area data points; Indicates the pitch angle data point In local area The fitted value of the data points; Indicates the pitch angle data point In local area The fitting error of the data points is Indicates the pitch angle data point The average fitting error of all data points in the local range. When the average fitting error is larger, the pitch angle data point The greater the degree of irregular fluctuation in the local area, the greater the pitch angle data point The greater the irregular volatility.

[0059] It should be noted that the present invention fits the data points within the local range of the single-dimensional posture data points, and uses the fitting error to represent the irregular volatility of the single-dimensional posture data points. This is because the more irregular the changes in the data sequence, the greater the fitting error when performing curve fitting.

[0060] Similarly, calculate the pitch angle sequence Irregular fluctuations of data points and yaw angle sequence Irregular fluctuations of data points and roll angle series The irregular fluctuation of each data point in .

[0061] S202: Determine the noise possibility of the pitch angle data point according to the difference between the irregular fluctuation of the pitch angle data point and the irregular fluctuation of the yaw angle data point and the roll angle data point at the same time.

[0062] It should be noted that, due to the fact that there may be some normal attitude changes during the operation of the drone, which may cause some irregular or drastic changes in the attitude data of the drone, such as fancy operations of the drone, resulting in large irregular volatility of the data points. If the irregular volatility is directly used as the possibility of noise, the normal data changes may be mistaken for noise. Therefore, the present invention performs credibility analysis on the irregular volatility of the data points: Affected by the airflow during the flight, the time when the attitude data of each dimension of the drone appears noise is similar, while in normal attitude changes, there is a high probability that only one dimension of data changes irregularly at the same time, that is, if it is noise, then the irregular volatility of the pitch angle, yaw angle, and roll angle data points at the same time is similar, while in normal attitude changes, the irregular volatility of the pitch angle, yaw angle, and roll angle data points at the same time is not similar. Therefore, the present invention uses the difference in the irregular volatility of attitude data points in multiple dimensions at the same time to obtain the credibility of the irregular volatility of the data points as the possibility of noise, thereby obtaining the noise possibility of the data points.

[0063] Specifically, the pitch angle data points The noise probability satisfies the expression:

[0064] ;

[0065] In the formula, Indicates the pitch angle data point The possibility of noise, Indicates the pitch angle data point Irregular volatility; , Respectively represent the data points with pitch angle Irregular fluctuations of yaw angle data points and roll angle data points at the same time; represents an exponential function with a natural constant as base; Indicates the pitch angle data point at the same time The fluctuation similarity with the yaw angle data point and the roll angle data point, that is, the pitch angle data point The irregular fluctuation of is used as the credibility of the noise possibility. Average irregular volatility The greater the difference, the smaller the fluctuation similarity, and the pitch angle data point The less confidence there is in the possibility that the irregular fluctuations are noise, the more likely the pitch angle data points are. The less noise is likely to occur.

[0066] Similarly, get the pitch angle sequence The noise probability of each data point in the yaw angle sequence The noise probability of each data point and the roll angle sequence The noise probability of each data point in .

[0067] S203, perform iterative SG filtering on the attitude data of the UAV, determine the preferred degree of each iteration based on the correlation between the noise possibility of each data point in the attitude data of the UAV and the filtering amount of each data point in each iteration process, set the iteration stop condition based on the preferred degree, and obtain the final filtering result of the attitude data of the UAV.

[0068] Iterative SG filtering is performed on the pitch angle sequence F, where the window width of the SG filter is , the polynomial order is . The implementers can set the , In the present invention , .

[0069] The first The result of the iteration is recorded as , is a sequence of the same length as F. Subtract it from the pitch angle sequence F to obtain the filtering amount of each data point in the pitch angle sequence F.

[0070] It should be noted that, when the filtering amount is larger, the degree of filtering of the data point is greater. For data points with greater noise possibility, a larger filtering amount is required to better denoise the data point, and the data points with smaller noise possibility require a smaller filtering amount to better retain the detailed information of the pitch angle sequence. That is, the higher the correlation between the filtering amount and the noise possibility in a certain iterative filtering, the better the filtering effect of the iteration. Therefore, the present invention determines the preferred degree of each iteration of iterative SG filtering according to the correlation between the filtering amount of the pitch angle data point and the noise possibility, so as to determine whether to stop the iteration later according to the preferred degree.

[0071] Specifically, a rectangular coordinate system is established with the noise possibility as the horizontal axis and the filtering amount as the vertical axis. According to the noise possibility and the filtering amount of each pitch angle data point in the pitch angle sequence F, each pitch angle data point in the pitch angle sequence F is plotted in the rectangular coordinate system to obtain a scatter plot.

[0072] The points in the scatter plot are linearly fitted using the least squares method to obtain a linear regression curve of the scatter plot. The linear regression curve can reflect the changing relationship between the filtering amount and the noise possibility. Therefore, the present invention determines the first value when iterating SG filtering on the pitch angle sequence F according to the linear regression curve. The optimality of the iteration:

[0073] ;

[0074] in, It means that the pitch angle sequence F is iterated by SG filtering. The optimization degree of the iteration, Indicates After iterations, according to the pitch angle sequence The average distance from all points to the linear regression curve in the scatter plot established by the noise probability and filtering amount of each pitch angle data point, is an exponential function with a natural constant as the base. The average distance reflects the correlation between the noise possibility and the filtering amount. When the average distance from all points in the scatter plot to the linear regression curve is smaller, it means that all points in the scatter plot are closer to the linear regression curve, and the filtering amount of the pitch angle data point is more correlated with the noise possibility. When the average distance from all points in the scatter plot to the linear regression curve is larger, the correlation between the filtering amount of the pitch angle data point and the noise possibility is worse.

[0075] It should be noted that, in order to prevent under-filtering or over-filtering, the present invention sets the iterative stop condition of the iterative SG filtering according to the optimization degree of each iteration of the iterative SG filtering.

[0076] Specifically, in one embodiment, the iterative stopping condition of iterative SG filtering is:

[0077] In response to the iterative SG filtering, the preference level of a certain iteration in the iterative process is greater than the preset stop threshold. , stop the iteration, the filtering effect at this time has met the standard, and the result of this iteration is used as the final filtering result of the pitch angle sequence F. Among them, the stop threshold The specific value of can be set according to the actual application scenario and requirements. Since the value range of the preferred degree is [0,1], the stop threshold Used to judge the degree of preference, the stop threshold The value range of is also [0,1]. In this embodiment, the stopping threshold Set to 0.8.

[0078] In another embodiment, the iterative stopping condition of iterative SG filtering is:

[0079] In response to the iterative SG filtering, the preference level of a certain iteration in the iterative process is greater than the preset stop threshold. , and the optimization degree of this iteration is less than that of the previous iteration, the iteration is stopped, and the result of the previous iteration is used as the final filtering result of the pitch angle sequence F. Among them, the stop threshold The specific value of can be set according to the actual application scenario and requirements. Since the value range of the optimization degree is [0,1], the stop threshold Used to judge the degree of preference, the stop threshold The value range of is also [0,1]. In this embodiment, the stop threshold T is set to 0.8.

[0080] Similarly, iterative SG filtering is performed on the yaw angle sequence P and the roll angle sequence G respectively to obtain a final filtering result of the yaw angle sequence P and a final filtering result of the roll angle sequence G.

[0081] S3. Monitor the flight status of the UAV based on the final filtering result of the UAV's attitude data.

[0082] After filtering is completed, in order to obtain direct angle data, the present invention will The filtering result of the pitch angle data within the range is restored to Within the range, The filtering results of the yaw angle data and the rolling angle data within the range are restored to within the range.

[0083] Specifically, in one embodiment, when step S1 adopts , , When the pitch angle, yaw angle and roll angle are preprocessed, the restoration process of the pitch angle is: ,in is the pitch angle after filtering, To filter the pre-processed pitch angle data, is the inverse sine function; the restoration process of the yaw angle is: ,in is the yaw angle after filtering, The data is filtered for the preprocessed yaw angle data; the restoration process of the roll angle is: ,in is the roll angle after filtering, The data is obtained by filtering the pre-processed roll angle data.

[0084] In another embodiment, when step S1 adopts , , When the pitch angle, yaw angle and roll angle are preprocessed, the restoration process of the pitch angle is: ,in is the pitch angle after filtering, The data is filtered for the preprocessed pitch angle data; the restoration process of the yaw angle is: ,in is the yaw angle after filtering, The data is filtered for the preprocessed yaw angle data; the restoration process of the roll angle is: ,in is the roll angle after filtering, The data is obtained by filtering the pre-processed roll angle data.

[0085] In other embodiments, the implementer may perform the pre-processing of the pitch angle, yaw angle and roll angle according to the specific pre-processing process of the pitch angle, yaw angle and roll angle. The filtered results of the pitch angle data, yaw angle data, and roll angle data within the range are restored to obtain The pitch angle of the filtered range, The yaw angle and roll angle filtered within the range.

[0086] The flight status of the UAV can be accurately monitored based on the pitch angle, yaw angle and roll angle obtained through filtering.

[0087] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0088] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for monitoring the flight status of an air-to-ground integrated unmanned aerial vehicle, characterized in that: include: Collect the attitude data of the UAV, including pitch angle sequence, yaw angle sequence and roll angle sequence; Performing curve fitting on data points in a local range of each pitch angle data point in the pitch angle sequence, and determining the irregular fluctuation of each pitch angle data point according to the fitting error; Determine the noise possibility of the pitch angle data point according to the difference between the irregular fluctuation of the pitch angle data point and the irregular fluctuation of the yaw angle data point and the roll angle data point at the same time; Iterative SG filtering is performed on the pitch angle sequence, and the degree of optimization of each iteration is determined according to the correlation between the noise possibility of each pitch angle data point in the pitch angle sequence and the filtering amount of each pitch angle data point in each iteration process; In response to the preference degree being greater than a preset stop threshold, the iteration is stopped to obtain a final filtering result of the pitch angle sequence; The flight status of the UAV is monitored based on the final filtering results of the pitch angle sequence, yaw angle sequence and roll angle sequence.

2. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 1, characterized in that: The elements in the pitch angle sequence, yaw angle sequence and roll angle sequence are pre-processed data of the pitch angle, yaw angle and roll angle. within the range.

3. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 1, characterized in that: The irregular fluctuation of the pitch angle data points satisfies the expression: ； In the formula, Indicates the pitch angle data point The irregular fluctuations of is the first Pitch angle data points; Indicates the pitch angle data point The number of data points in the local range; Indicates the pitch angle data point In local area data points; Indicates the pitch angle data point In local area The fitted value of the data points; Indicates the pitch angle data point In local area The fitting error of the data points; is the absolute value symbol.

4. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 1, characterized in that: The noise probability of the pitch angle data point satisfies the expression: ； In the formula, Indicates the pitch angle data point The possibility of noise, Indicates the pitch angle data point Irregular volatility; , Respectively represent the pitch angle data points Irregular fluctuations of yaw angle data points and roll angle data points at the same time; Represents an exponential function with a natural constant as its base.

5. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 1, characterized in that: The method for obtaining the filtering amount is: The first The result of the iteration is recorded as , convert the sequence Subtract the pitch angle sequence from the pitch angle sequence to obtain the first The amount of filtering for each iteration.

6. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 1, characterized in that: The step of determining the optimization degree of each iteration according to the correlation between the noise possibility of each pitch angle data point in the pitch angle sequence and the filtering amount of each pitch angle data point in each iteration process comprises: A rectangular coordinate system is established with the noise possibility as the horizontal axis and the filtering amount as the vertical axis, and a scatter plot is drawn according to the noise possibility and filtering amount of each pitch angle data point in the pitch angle sequence; Linear fitting is performed on the points in the scatter plot to obtain a linear regression curve; the degree of preference for this iteration is determined according to the average distance from each point in the scatter plot to the linear regression curve, and the average distance reflects the correlation between the noise possibility and the filtering amount.

7. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 6, characterized in that: The preference degree satisfies the expression: ； in, Indicates the first The optimization degree of the iteration, Indicates After iterations, the average distance from all points to the linear regression curve in the scatter plot established based on the noise probability of each pitch angle data point in the pitch angle sequence and the filtering amount is: is an exponential function with a natural constant as its base.

8. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 2, characterized in that: Also includes: After filtering is completed, the final filtering result of the pitch angle sequence is restored to Within the range, restore the final filtering results of the yaw angle data sequence and the final filtering results of the roll angle sequence to within the range.

9. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 8, characterized in that: The pre-processing process is as follows: The preprocessing process of the pitch angle is: ,in is the pitch angle obtained, is the preprocessed pitch angle data, is a sine function; The preprocessing process of the yaw angle is: ,in is the collected yaw angle, is the preprocessed pitch angle data; The preprocessing process of the roll angle is: ,in is the collected roll angle, is the preprocessed roll angle data.

10. The method for monitoring the flight status of an air-to-ground integrated UAV according to claim 9, characterized in that: The restoration process is: The restoration process of the pitch angle is: ,in is the pitch angle after filtering, is the final filtering result of the preprocessed pitch angle data, is the inverse sine function; The restoration process of the yaw angle is: ,in is the yaw angle after filtering, is the final filtering result of the preprocessed yaw angle data; The process of restoring the roll angle is: ,in is the roll angle after filtering, is the final filtering result of the preprocessed roll angle data.

Images