A method, program product, device, and medium for detecting the consistency of observation results

Abstract

The present application provides a method, a program product, a device and a medium for detecting the consistency of observation results. The method includes: during the operation of a movable platform, obtaining multiple redundant observation results of the same observable quantity by the movable platform; determining the test statistic and the rejection region of each observation result in the hypothesis testing theory; the null hypothesis in the hypothesis testing theory is that all observation results are normal and follow a preset distribution; the alternative hypothesis is that there are abnormal observation results deviating from the preset distribution; for each observation result, determining the observation result whose test statistic falls into the rejection region as an abnormal observation result, and determining the observation result whose test statistic does not fall into the rejection region as passing the consistency test. Thus, abnormal observation results and normal observation results are distinguished in real time. Through automated data collection and processing, the work efficiency is improved and human errors are reduced. It also meets the detection requirements of various sensors and different environments, ensuring the long-term stable operation of the movable platform.

Description

Technical Field

[0001] This application relates to the technical field of mobile platforms. Specifically, it relates to a method, program product, device, and medium for detecting the consistency of observation results. Background Art

[0002] The precise positioning and navigation of a mobile platform rely on various sensors carried by itself to sense the surrounding environment and rigorous algorithm processing. Currently, in order to improve the reliability of observation results, a mobile platform can perform redundant observations on the same observed quantity through multiple sensors and / or multiple processing algorithms to obtain multiple redundant observation results, and perform consistency detection on these redundant observation results. In the related art, the consistency detection of multiple observation results mostly relies on offline manual comparison and basic data analysis, which has problems of low efficiency and poor accuracy. Moreover, it is difficult to simultaneously perform consistency detection on the redundant observation results of multiple observed quantities, and it is unable to effectively cope with the impact of environmental changes on the performance of sensors. Summary of the Invention

[0003] The purpose of the embodiments of this application is to provide a method, program product, device, and medium for consistency detection to solve at least one of the above technical problems.

[0004] The first aspect of the embodiments of this application provides a method for detecting the consistency of observation results, which is used to detect the consistency of multiple redundant observation results of the same observed quantity; the method includes:

[0005] During the operation of the mobile platform, obtain multiple redundant observation results of the mobile platform for the same observed quantity;

[0006] Determine the test statistic and rejection region of each observation result in the hypothesis testing theory; wherein, the null hypothesis in the hypothesis testing theory is: all observation results are normal and follow a preset distribution; the alternative hypothesis in the hypothesis testing theory is: there are abnormal observation results deviating from the preset distribution;

[0007] For each observation result, determine that the observation result whose test statistic falls into the rejection region is an abnormal observation result, and determine that the observation result whose test statistic does not fall into the rejection region passes the consistency detection.

[0008] In the above implementation process, the hypothesis testing theory is utilized to perform consistency detection on multiple observation results obtained in real time during the operation of the mobile platform, thereby distinguishing abnormal observation results from normal observation results in real time, effectively improving the reliability and stability of the observation results. Through the automated data acquisition and processing process, manual intervention is greatly reduced, work efficiency is improved, human errors are reduced, and the accuracy of consistency detection is enhanced. Additionally, the consistency detection of redundant observation results of multiple observables can be achieved simultaneously through an automatic algorithm. Moreover, since this method is applicable to the operation process of the mobile platform, it can meet the detection requirements of various types of sensors and different environmental conditions, can promptly identify sensors with poor environmental robustness, discover problems in a timely manner and make corrections, ensuring the long-term stable operation of the mobile platform.

[0009] Further, before determining the test statistic and rejection region of each of the observation results in the hypothesis testing theory, the method further includes:

[0010] Obtain the observation differences between every two of the multiple observation results to obtain multiple such observation differences;

[0011] Determine that at least one of the observation differences is less than a preset difference threshold.

[0012] In the above implementation process, performing consistency testing using the hypothesis testing theory only when it is determined that at least one observation difference is less than the difference threshold can ensure that the consistency test result is valid and credible.

[0013] Further, the method further includes:

[0014] If it is determined that all the observation differences are greater than the difference threshold, determine that all the observation results are invalid.

[0015] In the above implementation process, by pairwise comparing the observation differences between the observation results and performing consistency detection using the hypothesis testing theory only when at least one observation difference is less than the difference threshold, and determining that all observation results are invalid when all observation differences are greater than the difference threshold, the effectiveness of the consistency detection result can be ensured.

[0016] Further, the observable quantity includes the attitude of the mobile platform; the observation result is the attitude angle; the multiple attitude angles include any combination of the attitude angles obtained by double-antenna direction finding carried by the mobile platform, the attitude angles obtained by visual positioning, the attitude angles measured by a magnetic compass carried by the mobile platform, and the attitude angles obtained by solving based on speed observations; the obtaining of the observation differences between every two of the multiple observation results includes:

[0017] Convert each attitude angle into a two-dimensional unit vector respectively;

[0018] For every two attitude angles, calculate the dot product of the corresponding two two-dimensional unit vectors, and determine the included angle between the two two-dimensional unit vectors based on the dot product; wherein, the included angle is the observed difference between the two attitude angles.

[0019] In the above implementation process, by converting the attitude angle into a two-dimensional unit vector, using the relationship between the dot product and the included angle between vectors to obtain the observed difference between two attitude angles, by comparing the angular differences between attitude angles pairwise, and using the hypothesis testing theory to perform consistency detection on all attitude angles when at least one angular difference is less than the angle threshold, the effectiveness of the consistency detection result can be ensured.

[0020] Further, the movable platform is also equipped with a gyroscope; the method further includes:

[0021] If it is determined that all the observed differences are greater than the difference threshold, determine that all the attitude angles are invalid, and use the gyroscope to determine the target attitude angle of the unmanned aerial vehicle.

[0022] In the above implementation process, if it is determined that the angular differences between all attitude angles are greater than the angle threshold, then determine that all attitude angles are invalid, and use the angular velocity sensed by the gyroscope to integrate over time to obtain the target attitude angle, thereby ensuring the observation accuracy of the observable quantity of the attitude of the movable platform and providing a reliable target attitude angle for subsequent positioning and navigation.

[0023] Further, the determination of the test statistic and the rejection region of each of the observed results in the hypothesis testing theory includes:

[0024] Determine the expected value based on multiple observed results;

[0025] For each observed result, calculate the test statistic of the observed result based on the value of the observed result, the standard deviation, the number of observations, and the expected value;

[0026] Based on the preset significance level and the redundancy quantity of the observed result, determine the critical value of the rejection region to obtain the rejection region.

[0027] In the above implementation process, use multiple observed results of the same observable quantity to determine the expected value, determine the rejection region based on the significance level and the redundancy quantity, and then evaluate whether the observed result deviates from the expected value based on whether the test statistic of each observed result falls into the rejection region, so as to find out the abnormal observed results.

[0028] Further, the movable platform is equipped with sensors for observing the observed quantities; a plurality of the observation results include sensor observation results obtained by observing using the sensors; the method further includes:

[0029] If the sensor is a first type of sensor, determine the preset standard deviation of the sensor as the standard deviation of the sensor observation result;

[0030] If the sensor is a second type of sensor, determine the real-time sample standard deviation of the sensor as the standard deviation of the sensor observation result; wherein, the environmental robustness of the first type of sensor is better than that of the second type of sensor.

[0031] In the above implementation process, different standard deviations of the sensor observation results are determined based on the environmental robustness of the sensors, balancing the efficiency and accuracy of the consistency detection.

[0032] A second aspect of the embodiments of the present application provides a computer program product, the computer program product includes a computer program, and when the computer program is executed by a processor, it implements the method according to any one of the first aspect.

[0033] A third aspect of the embodiments of the present application provides an electronic device, the electronic device includes:

[0034] A processor;

[0035] A memory for storing instructions executable by the processor;

[0036] Wherein, when the processor calls the executable instructions, it implements the operations of the method according to any one of the first aspect.

[0037] A fourth aspect of the embodiments of the present application provides a computer-readable storage medium, on which computer instructions are stored, and when the computer instructions are executed by a processor, it implements the steps of the method according to any one of the first aspect. Description of the Drawings

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required to be used in the embodiments of the present application. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0039] Figure 1 It is a schematic flowchart of a method for detecting the consistency of observation results provided by an embodiment of the present application;

[0040] Figure 2Schematic flowchart of another method for detecting the consistency of observation results provided by an embodiment of the present application;

[0041] Figure 3 Schematic flowchart of another method for detecting the consistency of observation results provided by an embodiment of the present application;

[0042] Figure 4 Schematic flowchart of another method for detecting the consistency of observation results provided by an embodiment of the present application;

[0043] Figure 5 Hardware structure diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0044] Next, the technical solutions in the embodiments of the present application will be described with reference to the accompanying drawings in the embodiments of the present application.

[0045] It should be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present application, the terms "first", "second", etc. are only used for descriptive distinction and cannot be construed as indicating or implying relative importance.

[0046] A mobile platform refers to any device that can move, which may include but is not limited to land vehicles, water vehicles, air vehicles, and other types of motorized carriers. As an example, the mobile platform can be an unmanned aerial vehicle (UAV), an unmanned vehicle, a robot, etc. A variety of sensors are usually mounted on the mobile platform to sense the surrounding environment, thereby obtaining observation results. The mobile platform can perform redundant observations on the same observable quantity to improve the reliability of the observation results.

[0047] Specifically, multiple redundant observation results of the same observable quantity need to pass consistency detection to ensure the credibility of the observation results. In related technologies, the consistency detection of multiple redundant observation results in a mobile platform mostly relies on offline manual comparison and basic data analysis. Specifically, multiple redundant observation results of the same observable quantity are obtained when the mobile platform is stationary, and then whether the multiple observation results meet the consistency requirements is manually compared. However, the detection method of manual comparison has problems of low efficiency and poor accuracy. Furthermore, since some observation results are sensed by sensors carried by the mobile platform, and the sensors may be affected by complex environments, resulting in fluctuations in the observation results. The observation results sensed by the sensors pass the consistency detection under the condition that the mobile platform is stationary, which does not mean that they can also pass the consistency detection in a complex and fluctuating environment. Therefore, offline manual comparison is difficult to meet the detection requirements under different usage conditions. Moreover, manual comparison can only detect the consistency of a single sensor or a single observable quantity, and cannot perform the consistency detection of redundant observation results of multiple observable quantities simultaneously.

[0048] To solve at least one of the above technical problems, the present application provides a method for consistency detection of observation results. This method is used to perform online consistency detection of multiple redundant observation results of the same observable quantity. The method can be executed by the mobile platform or by other computing devices such as a server or a control device that is communicatively connected to the mobile platform. The present application does not limit the execution entity of the method. Refer to Figure 1 , the consistency detection method includes Step 110 - Step 130.

[0049] Step 110: During the operation of the mobile platform, obtain multiple redundant observation results of the same observable quantity by the mobile platform.

[0050] Different from the above-mentioned consistency detection scheme in the case where the mobile platform is stationary, the execution timing of the consistency detection method provided in this embodiment is during the operation of the mobile platform. For example, if the mobile platform is an aircraft, then the method can be triggered and executed during the power-on, take-off, or flight of the aircraft. If the mobile platform is a robot, then the method can be triggered and executed when the robot starts to move.

[0051] An observed quantity refers to a specific physical quantity, state, or characteristic that can be acquired and recorded. An observation result refers to the quantified result of the observed quantity. The observed quantity may refer to the state quantity of the mobile platform itself during the operation process, such as speed, acceleration, and pose, etc.; it may also refer to the characteristic quantity of the environment where the mobile platform is operating, such as temperature, humidity, obstacle distance, etc. The observation result can be obtained in different ways. As an example, the observation result can be sensed based on the sensors mounted on the mobile platform. For example, the speed data sensed by a speed sensor, the attitude data sensed by a magnetic compass, the depth data sensed by a lidar, and so on. As another example, the observation result can be obtained by processing and calculating the operation data of the mobile platform through an algorithm. For example, the heading data calculated using the moving speed of the mobile platform, and so on. The present application does not limit the source of the observation result. Exemplarily, the consistency detection method provided in this embodiment can be executed periodically. The execution period can match the sensing period of the sensor and / or the calculation period of the algorithm. Then the multiple redundant observation results obtained in step 110 refer to the observation results obtained in real time by the mobile platform during the operation process.

[0052] Exemplarily, multiple observation results obtained for the same observed quantity can be stored in the form of an array. Specifically, each array corresponds to an observed quantity. The array includes multiple elements, and each element is an observation result of the observed quantity corresponding to the array. If there are observation results of multiple observed quantities of the mobile platform that need to be subjected to consistency detection simultaneously, then an array can be set for each observed quantity, and each array stores multiple observation results corresponding to the respective observed quantity. Subsequently, the observation results in the multiple arrays can be read through multiple parallel threads or processes, and the consistency detection of the observation results of multiple observed quantities can be performed simultaneously.

[0053] As an example, the array stores the latest obtained observation results for the same observed quantity in the current execution period. As another example, each observation result stored in the array is an observation result after noise reduction processing to reduce the interference of environmental noise on the consistency detection.

[0054] Step 120: Determine the test statistic and the rejection region of each of the observation results in the hypothesis testing theory; wherein, the null hypothesis in the hypothesis testing theory is: all the observation results are normal and follow a preset distribution; the alternative hypothesis in the hypothesis testing theory is: there are abnormal observation results that deviate from the preset distribution.

[0055] Step 130: For each observation result, determine that the observation result for which the test statistic falls into the rejection region is an abnormal observation result, and determine that the observation result for which the test statistic does not fall into the rejection region passes the consistency detection.

[0056] Hypothesis Theory is a method in statistics for determining whether sample data supports the null hypothesis H0. It can be used to determine whether the differences between samples and between samples and the population are caused by sampling errors or essential differences. By comparing the deviation between sample data and the expected value, it evaluates whether the deviation is large enough to obtain the conclusion of rejecting the null hypothesis H0 and accepting the alternative hypothesis H1, or the conclusion of being unable to reject the null hypothesis H0.

[0057] In this embodiment, Hypothesis Theory is used to verify the consistency of multiple observation results of the same observable. First, the null hypothesis H0 in Hypothesis Theory is set as: all observation results of the same observable are normal and follow a preset distribution. The preset distribution may include, but is not limited to, the normal distribution. And the alternative hypothesis H1 in Hypothesis Theory can be set as: there are abnormal observation results that deviate from the preset distribution. Subsequently, the deviation between each observation result and the expected value is compared. If it is evaluated that the deviation corresponding to one or some observation results is large enough, the null hypothesis H0 can be rejected and the alternative hypothesis H1 can be accepted. That is, it is determined that the observation results with large enough deviation deviate from the preset distribution and are abnormal observation results, and the abnormal observation results do not pass the consistency test. If it is evaluated that the deviation corresponding to one or some observation results is small, the null hypothesis H0 cannot be rejected. That is, it is determined that the observation results with small deviation follow the preset distribution and are normal observation results, and the normal observation results pass the consistency test.

[0058] Specifically, when comparing the deviation between each observation result and the expected value and evaluating whether the deviation is large enough, the test statistic and the corresponding rejection region of each observation result can be determined. Subsequently, for each observation result, it can be judged whether its corresponding test statistic falls into the corresponding rejection region. If the test statistic of the observation result falls into the rejection region, it indicates that the observation result does not support the establishment of the null hypothesis H0. Therefore, it can be determined that the observation result is an abnormal observation result that deviates from the preset distribution. If the test statistic of the observation result does not fall into the rejection region, that is, it falls into the acceptance region, it indicates that the observation result supports the establishment of the null hypothesis H0. Therefore, it can be determined that the observation result passes the consistency test and is a reliable normal observation result that follows the preset distribution.

[0059] Optionally, after determining the abnormal observation results and normal observation results from all the observation results, the abnormal observation results that do not pass the consistency test can be excluded, and the target observation result of this observable can be determined based on the observation results that pass the consistency test. For example, the average value or weighted average value of all the observation results that pass the consistency test can be determined as the target observation result of this observable.

[0060] It can be known that a method for detecting the consistency of observation results provided by the present application uses the hypothesis testing theory to detect the consistency of multiple observation results obtained in real time during the operation of a movable platform, so as to distinguish abnormal observation results from normal observation results in real time, which can effectively improve the reliability and stability of the observation results. Through the automated data collection and processing process, manual intervention is greatly reduced, the work efficiency is improved, and the human error is reduced, thereby improving the accuracy of the consistency detection. And through the automatic algorithm, the consistency detection of redundant observation results of multiple observables can be realized simultaneously. In addition, since this method is applicable to the operation process of the movable platform, it can meet the detection requirements of various types of sensors and different environmental conditions, can timely identify sensors with poor environmental robustness, discover problems in time and correct them, and ensure the long-term stable operation of the movable platform.

[0061] The following will introduce steps 110-step 130 in detail.

[0062] According to some embodiments of the present application, after performing step 110 and before performing step 120, the method may further include steps 210-step 220 as Figure 2 shown.

[0063] Step 210: Obtain the observation differences between every two of the multiple observation results to obtain a plurality of the observation differences;

[0064] Step 220: Determine that at least one of the observation differences is less than a preset difference threshold.

[0065] Exemplarily, if there are N observation results for the same observable, and the observation differences between every two observation results are compared, then a total of observation differences can be obtained. And among all the obtained observation differences, if at least one observation difference is less than the preset difference threshold, then continue to execute steps 120-130.

[0066] It can be understood that the existence of at least one observation difference less than the difference threshold indicates that there are at least two observation results that are relatively close among all the observation results. This is a judgment on whether it is effective to use the hypothesis testing theory for consistency testing, and it is also a preliminary judgment on the consistency of all the observation results. Using the hypothesis testing theory for consistency testing when it is determined that at least one observation difference is less than the difference threshold can ensure that the consistency test result is valid and credible. Optionally, the difference threshold may be a chi-square distribution critical value.

[0067] It should be noted that among all the obtained observation differences, even if there are some observation differences greater than the difference threshold, in the subsequent consistency detection, the consistency detection will still be performed on all the observation results together, and the two observation results corresponding to the observation differences greater than the difference threshold will not be excluded.

[0068] Furthermore, if it is determined that all the observation differences are greater than the difference threshold, it indicates that there are significant differences among all the observation results. It can be understood that the role of the hypothesis testing theory in this solution is to test whether multiple observation results of the same observable quantity follow a preset distribution or there are abnormal data deviating from the preset distribution. However, if there are significant differences among all the observation results, then these observation results do not follow the preset distribution. At this time, even if the hypothesis testing theory is used to perform consistency detection to identify normal observation results, they are not credible. Therefore, if it is determined that all the observation differences are greater than the difference threshold, it can be determined that all the observation results are invalid. At this time, the invalid observation results should not be used to determine the target observation result of the observable quantity.

[0069] It can be seen that by pairwise comparing the observation differences between the observation results and then using the hypothesis testing theory for consistency detection when at least one observation difference is less than the difference threshold, and determining that all the observation results are invalid when all the observation differences are greater than the difference threshold, the effectiveness of the consistency detection result can be ensured.

[0070] The following takes the attitude of a mobile platform as an example of the observable quantity for expansion. The attitude of the mobile platform can be represented by any one of the attitude angles such as pitch angle, roll angle, and yaw angle, or a combination of multiple attitude angles. If the observable quantity includes the attitude of the mobile platform, then the observation result can be one or more of the above attitude angles. Each attitude angle can be measured in multiple ways, thus obtaining redundant attitude angles. For example, multiple redundant yaw angles, multiple redundant pitch angles, etc. can be obtained. The multiple redundant attitude angles include any multiple of the attitude angles obtained based on the dual-antenna direction finding carried by the mobile platform, the attitude angles obtained based on visual positioning, the attitude angles measured based on the magnetic compass carried by the mobile platform, and the attitude angles obtained by solving based on speed observation. Based on this, obtaining the observation differences between every two observation results in step 210 can specifically include steps 310 - 320 as Figure 3 shown in.

[0071] Step 310: Convert each attitude angle into a two-dimensional unit vector.

[0072] A two-dimensional unit vector refers to a two-dimensional vector with a modulus of 1. The process of converting the attitude angle into a two-dimensional unit vector can refer to the related technology, and this embodiment will not expand it here.

[0073] Step 320: For every two pose angles, calculate the dot product of the corresponding two two-dimensional unit vectors, and determine the included angle between the two two-dimensional unit vectors based on the dot product; wherein, the included angle is the observed difference between the two pose angles.

[0074] Exemplarily, for every two pose angles, obtain the two-dimensional unit vectors corresponding to the two pose angles respectively, and calculate the dot product of the corresponding two two-dimensional unit vectors. For example, the two-dimensional unit vector corresponding to the pose angle γ , and the two-dimensional unit vector corresponding to the pose angle β , then the dot product of the two two-dimensional unit vectors can be expressed as:

[0075]

[0076] where is the modulus length of the two-dimensional unit vector , and its value is 1; is the modulus length of the two-dimensional unit vector , and its value is 1; δ angle is the included angle between the two-dimensional unit vector and the two-dimensional unit vector , that is, the observed difference between the pose angle γ and the pose angle β, and also the angular difference between the pose angle γ and the pose angle β. It can be seen that the included angle δ angle can be expressed as:

[0077]

[0078] where the LIMIT function is used to limit the input value range to [-1, 1] to avoid illegal input caused by numerical errors. The finally output included angle δ angle has a value range of [0, π], which is applicable to a variety of geometric calculation scenarios.

[0079] It can be seen that in this embodiment, by converting the pose angle into a two-dimensional unit vector, using the relationship between the dot product and the included angle between vectors to obtain the observed difference between two pose angles, by comparing the angular differences between pose angles pairwise, and using the hypothesis testing theory to perform consistency detection on all pose angles when at least one angular difference is less than the angle threshold, the effectiveness of the consistency detection result can be ensured.

[0080] After obtaining the angular differences between every two pose angles, it can be determined one by one whether each angular difference is greater than the angle threshold. If there is at least one angular difference δ angleIf it is less than a preset angular threshold, then the consistency of all attitude angles is detected using hypothesis testing theory. According to some embodiments of the present application, the movable platform is also equipped with a gyroscope, and the gyroscope can sense the angular velocity of the movable platform. Thus, on the basis of the above embodiments, if it is determined that all the observed differences are greater than the difference threshold, that is, all the angular differences δ angle are all greater than the preset angular threshold, it is determined that all the attitude angles are invalid. And the target attitude angle is obtained by integrating the angular velocity sensed by the gyroscope over time.

[0081] It can be seen that in this embodiment, if it is determined that the angular differences between all the attitude angles are all greater than the angular threshold, it is determined that all the attitude angles are invalid, and the target attitude angle is obtained by integrating the angular velocity sensed by the gyroscope over time, so as to ensure the observation accuracy of the observable quantity of the attitude of the movable platform and provide a reliable target attitude angle for subsequent positioning and navigation.

[0082] On the basis of any of the above embodiments, regarding determining the test statistic and the rejection region of each observation result in step 120, it may specifically include steps 121-step 123 as Figure 4 shown.

[0083] Step 121: Determine the expected value based on the multiple observation results.

[0084] The expected value refers to the true value of the observable quantity. However, in actual situations, since the true value of the observable quantity cannot be known, the average value or weighted average value of multiple observation results can be determined as the expected value.

[0085] Step 122: For each of the observation results, calculate the test statistic of the observation result based on the value, standard deviation, number of observations, and the expected value of the observation result.

[0086] Exemplarily, there are various types of test statistics, such as including but not limited to t-test, z-test, etc. Those skilled in the art can select the required type of test statistic according to actual needs and calculate the corresponding test statistic. For different types of test statistics, the test statistic of the observation result can be calculated based on the value of the observation result, the standard deviation of the observation result, the number of observations, and the expected value. Among them, the number of observations refers to the number of samples of the observable quantity, and the number of observations of different observation results can be the same or different.

[0087] Taking the z-test as an example, the test statistic Zi of each observation result can be expressed as:

[0088]

[0089] where is the value of the i-th observation result, μ is the expected value, is the standard deviation of the i-th observation result, and k is the number of observations.

[0090] Step 123: Determine the critical value of the rejection region based on a preset significance level and the redundancy quantity of the observation results, to obtain the rejection region.

[0091] Exemplarily, the significance level can be preset, for example, including but not limited to 0.05 or 0.01. The redundancy quantity of the observation results refers to how many observation results there are for a certain observed quantity. The redundancy quantity of the observation results can be used as the degrees of freedom, and the critical value of the rejection region can be found from the chi-square test table using the significance level and the degrees of freedom, to obtain the rejection region.

[0092] It can be known that in this embodiment, multiple observation results of the same observed quantity are used to determine the expected value, and the rejection region is determined based on the significance level and the redundancy quantity. Then, based on whether the test statistic of each observation result falls into the rejection region, it is evaluated whether the observation result deviates from the expected value, so as to find out the abnormal observation results.

[0093] In addition, as described above, the observation results can be obtained in different ways, including being sensed by a sensor, or being calculated by an algorithm using the operation data of a movable platform. In some embodiments, for multiple observation results of the same observed quantity, at least some of the observation results are obtained by observing the observed quantity using a sensor carried by a movable platform. This part of the observation results can be called sensor observation results. At this time, the standard deviation of the sensor observation results is related to the environmental robustness of the sensor. The environmental robustness of the sensor refers to whether the sensor can maintain its performance, stability, and reliability in the face of complex and changeable environments such as weather changes, terrain differences, and electromagnetic interference. In this embodiment, a sensor with better environmental robustness is called a first-class sensor, and a sensor with poorer environmental robustness is called a second-class sensor. It can be known that the first-class sensor and the second-class sensor are only distinguished based on the environmental robustness of the sensor, and do not specifically refer to a certain type of sensor.

[0094] Specifically, each sensor is calibrated with its own standard deviation when leaving the factory. If the sensor is a first-class sensor with better environmental robustness, the sensor observation results sensed by it are less affected by the environment, and the sensor observation results are relatively stable even in a complex and changeable environment. Therefore, when the movable platform is operating, the standard deviation of the sensor observation results will not deviate too far from the standard deviation calibrated when leaving the factory. Thus, when performing step 122, the preset standard deviation of the sensor when leaving the factory can be determined as the standard deviation of the sensor observation results. 。When performing step 122, the preset standard deviation can be directly read from the storage space, thereby improving the efficiency of consistency detection. If the sensor is a second type of sensor with poor environmental robustness, it means that the sensor observation results obtained by sensing are affected by the environment. At this time, during the operation of the mobile platform, the standard deviation of the sensor observation results may be quite different from the standard deviation calibrated when the sensor left the factory. Therefore, when performing step 122, the real-time sample standard deviation of the sensor can be determined as the standard deviation of the sensor observation results , thereby improving the accuracy of consistency checking.

[0095] It can be seen that in this embodiment, different standard deviations of sensor observation results are determined based on the environmental robustness of the sensor, balancing the efficiency and accuracy of consistency detection.

[0096] To better understand this solution, the heading attitude of an unmanned aerial vehicle is taken as an example of the observable quantity for illustration.

[0097] During the operation of the unmanned aerial vehicle, real-time heading attitude is required for positioning and navigation. The ways to obtain the heading attitude usually include but are not limited to the following: ① Using the RTK (Real-Time Kinematic) dual antennas carried by the unmanned aerial vehicle to measure the heading attitude of the unmanned aerial vehicle; ② Using the vision positioning sensing system carried by the unmanned aerial vehicle to measure the heading attitude; ③ Using the magnetic compass carried by the unmanned aerial vehicle to measure the heading attitude; ④ Based on the observation and calculation of the flight speed of the unmanned aerial vehicle to obtain the heading attitude. Using the above four acquisition methods, four observation results (four heading angles) of the heading attitude of the unmanned aerial vehicle at a certain moment can be obtained, that is, the redundancy quantity of the observation results is 4.

[0098] First, for each way of obtaining the heading attitude, noise reduction processing can be performed on the obtained heading angles. Taking the heading angles measured by the sensor as an example, k observation values of the heading attitude of the i-th sensor within a continuous period of time can be obtained, and the mean value of the k observation values is determined as the heading angle of the i-th sensor for the heading attitude , specifically expressed as:

[0099]

[0100] where x ij is the measurement value of the i-th sensor at the j-th time.

[0101] The obtained heading angle can be stored in the array corresponding to the heading attitude. In this embodiment, there are 4 observation results of the heading attitude, which are correspondingly stored in 4-bit elements of the data.

[0102] Subsequently, the observation differences between every two of the four heading angles can be evaluated, that is, whether the angular difference between every two heading angles is greater than a preset angular threshold. The specific process can refer to the above embodiments. If the angular differences between every two heading angles are all greater than the angular threshold, it indicates that the heading angles obtained by the four methods are quite different. In this case, it is possible that there are problems with all four acquisition methods, and the normal heading angle cannot be determined through consistency detection. Therefore, it is determined that the above four heading angles are all invalid, and the target heading angle is obtained by integrating the angular velocity sensed by the gyroscope on the unmanned aerial vehicle over time.

[0103] If there is at least one angular difference between two heading angles that is less than the angular threshold, then the hypothesis testing theory is used to perform consistency detection on the four heading angles. Specifically, for the sensor observation results, if the sensor is a first type of sensor with better environmental robustness, the preset standard deviation of the sensor is determined as the standard deviation of the sensor observation results . If the sensor is a second type of sensor with poor environmental robustness. Then calculate the real-time sample standard deviation of the sensor, and determine the real-time sample standard deviation as the standard deviation of the sensor observation results . The real-time sample standard deviation S i . The calculation formula is:

[0104]

[0105] Subsequently, calculate the test statistic of each observation result. The calculation process can refer to the above embodiments. And according to the preset significance level α and the redundancy number of the observation results (4 in this embodiment), look up the critical value Z α / 2 of the rejection region from the chi-square test table to obtain the rejection region.

[0106] For each observation result, if the test statistic of the observation result falls into the rejection region, that is , determine that this observation result is an abnormal observation result. If this observation result is sensed by the sensor, then it can be determined that this sensor is a faulty sensor. If the test statistic of the observation result does not fall into the rejection region, then determine that this observation result passes the consistency detection.

[0107] It can be seen that in this embodiment, for multiple redundant observation results obtained by different methods for the same observed quantity, first compare pairwise whether the observation differences between the observed quantities are within the difference threshold. If all the observation differences exceed the difference threshold, then all the obtained observation results are considered invalid, which is a preliminary judgment on the consistency of the observation results.

[0108] If there is at least one observed difference not exceeding the difference threshold, the hypothesis testing theory is used to perform consistency detection on each observation result, so as to distinguish abnormal observation results from normal observation results. If an abnormal observation result is sensed by a sensor, it can also be determined that the sensor has failed, so as to screen out the faulty sensor.

[0109] It can be seen that this embodiment realizes real-time consistency detection of multiple redundant observation results during the operation of the mobile platform, so as to distinguish abnormal observation results from normal observation results in real time, which can effectively improve the reliability and stability of the observation results. Through the automated data acquisition and processing process, manual intervention is greatly reduced, work efficiency is improved, human error is reduced, and the accuracy of consistency detection is improved. And through the automatic algorithm, the consistency detection of redundant observation results of multiple observation quantities can be realized simultaneously. In addition, since this method is applicable to the operation process of the mobile platform, it can meet the detection requirements of various types of sensors and different environmental conditions, can timely identify sensors with poor environmental robustness, discover problems in time and make corrections, and ensure the long-term stable operation of the mobile platform.

[0110] Based on the consistency detection method of an observation result described in any of the above embodiments, the present application also provides a computer program product, which includes one or more computer programs or instructions. The computer program or instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. When the computer program is executed by a processor, it realizes the consistency detection method of an observation result described in any of the above embodiments.

[0111] Based on the consistency detection method of an observation result described in any of the above embodiments, the present application also provides Figure 5 a schematic structural diagram of an electronic device as shown in Figure 5 , at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, there may also be other hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to realize the consistency detection method of an observation result described in any of the above embodiments.

[0112] As an example, the electronic device may be a mobile platform, such as including but not limited to drones, unmanned vehicles, robots, etc. As another example, the electronic device may be a computing device communicatively connected to the mobile platform, such as including but not limited to servers, central control devices, remote controllers, etc.

[0113] The present application also provides a computer storage medium storing a computer program, which when executed by a processor can be used to execute the method for detecting the consistency of observation results described in any of the above embodiments.

[0114] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0115] In addition, in each embodiment of the present application, the various functional modules can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0116] If the above functions are implemented in the form of software functional modules and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0117] The above are only embodiments of the present application and are not intended to limit the protection scope of the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application. It should be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0118] As described above, the above are only specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, and all should be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0119] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the presence of additional identical elements in the process, method, article or device including the said element.

Claims

1. A method for detecting consistency of observation results, characterized in that: The method is used to perform consistency detection on multiple redundant observation results of the same observation; the method comprises: During the operation of the movable platform, a plurality of redundant observation results of the movable platform on the same observation quantity are obtained; the movable platform is a drone; Determine the test statistic and rejection region of each observation result in the hypothesis test theory, including: determine the average or weighted average of multiple observation results as the expected value; for each observation result, calculate the test statistic of the observation result based on the value of the observation result, the standard deviation, the number of observations, and the expected value; determine the critical value of the rejection region based on the preset significance level and the redundant number of the observation result to obtain the rejection region; wherein the null hypothesis in the hypothesis test theory is: all observation results are normal and obey the preset distribution; the alternative hypothesis in the hypothesis test theory is: there are abnormal observation results that deviate from the preset distribution; For each of the observation results, determining that the observation result whose test statistic falls into the rejection domain is an abnormal observation result, and determining that the observation result whose test statistic does not fall into the rejection domain passes the consistency test; wherein, if the abnormal observation result is sensed by a sensor carried by the drone, determining that the sensor is a faulty sensor; A target observation result of the observation quantity is determined based on the observation result passing the consistency detection.

2. The method according to claim 1, characterized in that Before determining the test statistic and the rejection region of each of the observation results in the hypothesis testing theory, the method further comprises: Obtaining an observation difference between every two of the observation results in the plurality of observation results, to obtain a plurality of the observation differences; It is determined that at least one of the observed differences is less than a preset difference threshold.

3. The method according to claim 2, characterized in that The method further comprises: If it is determined that all the observed differences are greater than the difference threshold, it is determined that all the observed results are invalid.

4. The method according to claim 2 or 3, characterized in that: The observed quantity includes the attitude of the movable platform; the observation result is an attitude angle; the multiple attitude angles include any multiple of an attitude angle obtained based on dual antenna direction finding carried by the movable platform, an attitude angle obtained based on visual positioning, an attitude angle measured based on a magnetic compass carried by the movable platform, and an attitude angle obtained based on velocity observation and solution; the obtaining of the observation difference between each two of the multiple observation results includes: Convert each posture angle into a two-dimensional unit vector; For every two posture angles, the dot product of the corresponding two two-dimensional unit vectors is calculated, and the angle between the two two-dimensional unit vectors is determined based on the dot product; wherein the angle is the observed difference between the two posture angles.

5. The method according to claim 4, characterized in that The movable platform is also equipped with a gyroscope; the method further comprises: If it is determined that all of the observed differences are greater than the difference threshold, all of the attitude angles are determined to be invalid, and the gyroscope is used to determine the target attitude angle of the movable platform.

6. The method according to claim 1, characterized in that The movable platform is equipped with a sensor for observing the observed quantity; the plurality of observation results include sensor observation results obtained by observing with the sensor; the method further includes: If the sensor is a first type of sensor, determining a preset standard deviation of the sensor as the standard deviation of the sensor observation result; If the sensor is a second type of sensor, the real-time sample standard deviation of the sensor is determined to be the standard deviation of the sensor observation result; wherein the environmental robustness of the first type of sensor is better than the environmental robustness of the second type of sensor.

7. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

8. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing processor-executable instructions; Wherein, when the processor calls the executable instruction, it implements the operation of any method described in claims 1-6.

9. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed by a processor, the steps of any method described in claims 1-6 are implemented.

Images