An advanced receiver autonomous integrity monitoring method considering fault detection and exclusion

Abstract

The present application relates to a kind of advanced receiver autonomous integrity monitoring method considering fault detection and elimination, belong to satellite navigation receiver technical field.Compared with only considering the ARAIM method of fault detection, the ARAIM method considering fault detection and elimination disclosed in the present application can ensure to meet continuity requirement by executing fault elimination algorithm, thereby guaranteeing the continuity of navigation service.For future multi-constellation global navigation satellite system, the method has stronger applicability, and can also meet more stringent navigation performance requirements.Compared with existing ARAIM method considering fault detection and elimination, the ARAIM method considering fault detection and elimination disclosed in the present application proposes a continuity requirement allocation scheme by introducing key parameters.The scheme can maximize the integrity performance sacrifice while meeting the continuity requirement, thereby improving the overall navigation service performance.

Description

TECHNICAL FIELD

[0001] The present application relates to an advanced receiver autonomous integrity monitoring method considering fault detection and exclusion, and belongs to the technical field of satellite navigation receivers. BACKGROUND

[0002] Global Navigation Satellite System (GNSS) is widely used in civil aviation field, and has greatly promoted the development of navigation industry. Integrity is the top priority of GNSS application, and is directly related to the safety of life and property of users. GNSS integrity is a measure of the degree of confidence in the correctness of the information provided by the entire navigation system. In order to evaluate GNSS integrity and further ensure the reliability of positioning results, receiver autonomous integrity monitoring (RAIM) algorithm is a very recommended method, which can realize fault detection according to consistency test theory by using redundant measurement information.

[0003] At present, there are two RAIM algorithms that are concerned: least square residual RAIM algorithm for single fault condition and multi-hypothesis solution separation (MHSS) RAIM algorithm suitable for multi-fault condition. Based on the MHSS-RAIM algorithm, the Federal Aviation Administration of the United States proposed an advanced RAIM (ARAIM) algorithm, which has the potential to meet the integrity requirements of positioning accuracy for vertical navigation at 200 feet. However, as the development of RAIM algorithm, ARAIM algorithm is more dependent on redundant observation information than RAIM algorithm.

[0004] With the comprehensive development of American GPS, Russian GLONASS, European Galileo and Chinese Beidou satellite navigation system, more redundant dual-frequency ranging signals will be provided by multi-constellation GNSS in the future. At this time, GNSS applications involving ARAIM algorithm will also develop vigorously. However, due to the large number of satellites in multi-constellation GNSS, it is common to interrupt navigation tasks due to fault detection. Therefore, fault exclusion algorithm may need to be performed to ensure the continuity of navigation service. However, false exclusion may increase the integrity risk. In order to improve the overall navigation service performance, the integrity performance and continuity performance should be considered. SUMMARY

[0005] The technical problem to be solved by the present application is that the existing ARAIM method is difficult to guarantee the overall navigation service performance when considering fault detection and exclusion at the same time. An advanced receiver autonomous integrity monitoring method considering fault detection and exclusion is proposed.

[0006] The purpose of the present application is realized by the following technical scheme:

[0007] An advanced receiver autonomous integrity monitoring method considering fault detection and exclusion, comprising the following steps:

[0008] Step one, obtaining positioning estimation solution of linearized pseudo-range observation equation and error of positioning estimation solution based on multi-hypothesis solution separation method;

[0009] Step two, constructing fault detection test statistic and fault exclusion test statistic based on positioning estimation solution and error of positioning estimation solution obtained in step one;

[0010] Step three, establishing integrity risk equation and continuity risk equation according to fault detection test statistic constructed in step two in the case of only considering fault detection, and establishing integrity risk equation and continuity risk equation according to fault detection test statistic and fault exclusion test statistic constructed in step two in the case of considering fault detection and exclusion simultaneously;

[0011] Step four, performing integrity risk evaluation on integrity risk equation established in step three by using protection level;

[0012] Step five, performing continuity risk allocation on continuity risk equation established in step three by introducing key parameter λ, and calculating fault detection threshold and exclusion threshold;

[0013] Step six, determining specific value of key parameter λ in step five by numerical method, and determining protection level in step four according to fault detection threshold and exclusion threshold calculated in step five, to complete advanced receiver autonomous integrity monitoring in the case of considering fault detection and exclusion simultaneously.

[0014] In the step one, the linearized pseudo-range observation equation is:

[0015] The linearized pseudo-range observation equation with n measurement values and m state values is expressed in the northeast celestial coordinate system as:

[0016] z = Hx + v + f (26)

[0017] In the formula, z is n×1 dimensional pseudo-range offset vector from real position to linearization point; H is n×m dimensional geometric matrix from positioning domain to measurement domain; x is m×1 dimensional state vector containing receiver three-dimensional position and clock parameter; v = [v1, v2, …, v n ] T n×1 dimensional Gaussian observation noise vector with mean value 0 and covariance matrix Σ; f is n×1 dimensional fault vector;

[0018] In the step one, the positioning estimation solution of state vector x in the linearized pseudo-range observation equation and the error of positioning estimation solution are calculated by using weighted least square algorithm, and the positioning estimation solution includes full set positioning estimation solution and subset positioning estimation solution;

[0019] Let α be an m x 1 vector, which is used to extract a certain state of interest 'x' from the state vector ensemble x (e.g., the vertical direction is primarily considered in the civil aviation field):

[0020]

[0021] The weighted least squares solution of state x is obtained using all visible satellites and its positioning error ε0is:

[0022]

[0023] where,

[0024] W =∑ -1 is the weight matrix, and the ensemble positioning error ε0obeys a Gaussian distribution:

[0025]

[0026] Under a certain fault assumption, the matrix W is defined i which has the same dimension as W, but the diagonal elements corresponding to the rows of the faulty satellites are 0, so the weighted least squares solution of state x is obtained based on the subset of visible satellites and its positioning error ε i is:

[0027]

[0028] where,

[0029] The subset positioning error ε i obeys a Gaussian distribution:

[0030]

[0031] The solution separation is defined as the difference between the ensemble positioning estimation solution and the subset positioning estimation solution:

[0032]

[0033] Substitute equation (28) and equation (31) into equation (34), Δ i can be further expressed as:

[0034]

[0035] The solution separation Δ i obeys a Gaussian distribution:

[0036]

[0037] In step two, the fault detection test statistic constructed is:

[0038] In the case of no fault, the fault vector f is a zero vector, and Δ i obeys a Gaussian distribution with mean 0; however, when there is a fault, Δ i obeys a Gaussian distribution with non-zero mean. Therefore, Δ i is taken as the test statistic to achieve satellite fault detection by comparing with a detection threshold;

[0039]

[0040] where D denotes detection of fault, ND denotes no detection of fault; q i denotes the fault detection test statistic, T i denotes the corresponding detection threshold; i = 0, 1, …, h is a set of mutually exclusive and exhaustive fault hypotheses, and h denotes the number of fault hypotheses;

[0041] In the second step, the constructed fault exclusion test statistic is:

[0042] When a satellite fault is detected, fault exclusion is needed to ensure the continuity of the navigation system. A subset of all the satellites detected with fault ‘i’ is selected as the exclusion candidate, denoted as S i . Using each satellite subset in S i as the new full set of visible satellites, a fault detection test statistic is constructed using multi-hypothesis resolution to perform the second round of fault detection. When no fault is detected in a subset, it means that the fault exclusion is successful;

[0043]

[0044] where E denotes exclusion of fault, NE denotes no exclusion of fault; q i,j denotes the fault exclusion test statistic, T i,j denotes the corresponding exclusion threshold, j = 0, 1, …, h and j ≠ i;

[0045] In the third step, only in the case of fault detection, the integrity risk equation and the continuity risk equation are:

[0046] The integrity risk is equivalent to the probability of hazardous misleading information (HMI), which refers to the probability that the true position exceeds the maximum allowable position error boundary without being detected. Therefore, the integrity risk equation is expressed as:

[0047]

[0048] where ε0denotes the horizontal positioning error or the vertical positioning error; corresponds to denotes the horizontal alert limit or the vertical alert limit; P i denotes the prior probability of the occurrence of the i-th fault hypothesis;

[0049] The continuity risk is equivalent to the loss of continuity (LOC) probability, which is the probability of all events that cause the mission interruption, including the cases of no fault and fault detection, and other events that cause the loss of continuity. Therefore, the continuity risk equation is expressed as:

[0050]

[0051] The integrity risk equation and the continuity risk equation under the condition of fault detection and exclusion are:

[0052] For a complete description of the integrity risk under the condition of fault detection and exclusion, all events that produce dangerous misleading information need to be considered, including the dangerous misleading information existing in the full set of positioning estimation solutions and the dangerous misleading information existing in the subset of positioning estimation solutions after the exclusion of the faulty satellite under a certain fault hypothesis j. Therefore, the integrity risk equation under the condition of fault detection and exclusion is expressed as:

[0053]

[0054] In the formula, corresponding to the integrity risk generated by executing the fault exclusion algorithm;

[0055] At the same time, the continuity risk is used to describe the mission interruption caused by the detection of the fault but the failure to exclude the fault. Therefore, the continuity risk equation under the condition of fault detection and exclusion is expressed as:

[0056]

[0057] Compared with formula (40), formula (42) reduces the continuity risk by executing the fault exclusion algorithm. However, the sacrifice for this is represented as in formula (41) that is, the increase in the integrity risk;

[0058] In step four, the method for evaluating the integrity risk by introducing the protection level is:

[0059] In the process of evaluating the integrity risk, it is challenging to directly use formula (41) because the joint probability distribution of the positioning error is relatively complex and computationally intensive. Therefore, first, the upper bound of P HMI,FDE is derived:

[0060]

[0061] Under a certain fault hypothesis, the probability distribution of each term on the right side of formula (43) is known, which is expressed as:

[0062]

[0063] where Q(*) = 1 - Φ(*) and Φ(*) denotes the standard normal cumulative distribution function;

[0064] To meet the integrity requirement, a commonly used approach is to compute a position error bound called the protection level (PL) such that the actual position has a sufficiently large probability of falling within the error bound. The protection level can be computed by replacing I in equation (44) with PL and solving for P HMI,FDE obtained by restricting within the integrity requirement;

[0065] P HMI,FDE ≤ A + B + C + D + E ≤ I req,x P NM

[0066]

[0067]

[0068] where I req,x denotes the integrity requirement allocated to the direction of state x; P NM denotes the prior probability of a fault that rarely occurs and does not need to be monitored;

[0069] As can be seen from equation (45), given the alarm limit I and the true navigation condition, the protection level is only related to B and E, i.e., only related to the detection threshold T i and the exclusion threshold T j,i . It is worth noting that B and E are monotonic increasing functions of T i and T j,i , respectively;

[0070] In step five, during the continuity risk allocation process, the upper bound of P LOC,FDE is first derived:

[0071]

[0072] where the second term indicates that, in addition to other continuity losses P other , the continuity loss due to fault detection and exclusion under each fault assumption constitutes a significant portion of the continuity requirement. Therefore, C req P other can be equally allocated to each fault assumption:

[0073]

[0074] In fact, by introducing the parameter λ, C req,i can be further allocated to the fault detection and fault exclusion tests:

[0075] P{|q i |>T i | 0}P0≤C req,i,d =λ·C req,i (48)

[0076]

[0077] Then, the detection threshold and the exclusion threshold are determined according to the formula (48) and (49) as follows:

[0078]

[0079] In the formula, Q -1 (*) represents the anti-tail probability distribution function of the two-tailed standard normal distribution;

[0080] According to the formula (50), the detection threshold T i is a monotone decreasing function of λ, and the exclusion threshold T j,i is a monotone increasing function of λ. Reviewing the relationship between the protection level and T i and T j,i discussed in step four, it is further concluded that there is a certain λ that makes the protection level minimum;

[0081] In the step six, the method for determining the specific value of the key parameter λ is as follows:

[0082] In order to ensure that the determined parameters are more accurate, the integrity risk under the condition of different λ is calculated according to the formula (44), wherein λ is increased from 0 to 1 with a step of 0.01. Then, the λ corresponding to the minimum integrity risk is found and the above two steps are repeated in each epoch. Finally, according to the statistical results, the key parameter is determined to be 0.05 or 0.15;

[0083] In the step six, based on the determined key parameter λ, the protection level is calculated by using the bisection method according to the formula (45) under the condition of only considering fault detection and under the condition of considering fault detection and exclusion at the same time. The obtained protection level is compared with the alarm limit value.

[0084] Beneficial effects

[0085] 1. Compared with the ARAIM method considering only fault detection, the ARAIM method considering fault detection and exclusion disclosed in the application can ensure that the continuity requirement is met by executing the fault exclusion algorithm, thereby ensuring the continuity of navigation service. For future multi-constellation global navigation satellite systems, the method has stronger applicability and can also meet more stringent navigation performance requirements.

[0086] 2、Compared with the existing ARAIM method considering fault detection and exclusion, the ARAIM method considering fault detection and exclusion disclosed in the application proposes a continuity requirement allocation scheme by introducing a key parameter. The scheme can maximize the reduction of the integrity performance sacrifice while meeting the continuity requirement, thereby improving the overall navigation service performance.

[0087] 3、The method of the application is based on multi-hypothesis solution separation to obtain full set positioning estimation solution and subset positioning estimation solution of linearized pseudo-range observation equation and positioning error thereof; based on the obtained positioning estimation solution and positioning error, a fault detection and fault exclusion test statistic is constructed; the integrity risk and continuity risk under the condition of only considering fault detection and under the condition of considering fault detection and exclusion are compared; a protection level is introduced for integrity risk evaluation; a key parameter is introduced for continuity risk allocation, and a fault detection and exclusion threshold is calculated; the specific value of the key parameter is determined by numerical method, and the protection level under the condition of only considering fault detection and under the condition of considering fault detection and exclusion is calculated to evaluate the availability of the ARAIM method under different conditions. The key point and the point to be protected of the application are a key parameter introduced in the ARAIM method, a continuity requirement allocation scheme based on the parameter, and a determination method of the parameter. BRIEF DESCRIPTION OF DRAWINGS

[0088] Figure 1 Minimum integrity risk quantity changes with key parameter;

[0089] Figure 2 Relationship between integrity risk and key parameter;

[0090] Figure 3 Comparison of protection levels of different methods. DETAILED DESCRIPTION

[0091] In order to better illustrate the purpose, content and advantages of the application, the specific embodiments will be further described in detail in combination with the embodiments of the application and the drawings.

[0092] EMBODIMENT

[0093] In this embodiment, a high-precision T300 GNSS receiver is used to collect real satellite navigation files and observation data from 00:00:00 to 23:59:59 on February 13, 2022, and the collection frequency is 1 Hz, so a total of 86400 epochs are obtained. Based on the collected data, the MWORKS software is used to simulate the whole process of GNSS receiver assisted civil aviation aircraft navigation positioning and fault detection and exclusion using ARAIM algorithm, so as to test the performance of the ARAIM method considering fault detection and exclusion disclosed in the application, and compare it with the ARAIM method considering only fault detection.

[0094] The linearized pseudo-range observation equation in step one is:

[0095] The linearized pseudo-range observation equation with n measurements and m state values in the ENU coordinate system is:

[0096] z = Hx + v + f (51)

[0097] where z is the n x 1 pseudo-range offset vector from the true position to the linearization point; H is the n x m geometric matrix from the positioning domain to the measurement domain; x is the m x 1 state vector containing the three-dimensional position of the receiver and the clock parameter; v = [v1, v2,..., vn]T is the n x 1 Gaussian observation noise vector with mean 0 and covariance matrix Σ; f is the n x 1 fault vector. n ] T

[0098] In step one, the positioning estimation solution of the state vector x in the linearized pseudo-range observation equation and the error of the positioning estimation solution are calculated using the weighted least squares algorithm, and the positioning estimation solution includes the full set positioning estimation solution and the subset positioning estimation solution.

[0099] Let a be an m x 1 vector, which is used to extract a state of interest 'x' from the state vector full set x (for example, the vertical direction is primarily considered in the civil aviation field):

[0100]

[0101] The weighted least squares solution of the state x is obtained using all visible satellites and the positioning error ε0 of the positioning estimation solution is:

[0102]

[0103] where

[0104]

[0105] W = Σ -1 is the weighting matrix. The full set positioning error ε0 follows a Gaussian distribution:

[0106]

[0107] Under a certain fault assumption, the matrix W is defined i which has the same dimension as W, but the diagonal elements corresponding to the rows of the faulty satellites are 0. Therefore, the weighted least squares solution of the state x is obtained based on the subset visible satellites and the positioning error ε of the positioning estimation solution is: i

[0108] ​​

[0109] where,

[0110]

[0111] subset positioning error ε i obeys a Gaussian distribution:

[0112]

[0113] The solution separation is defined as the difference between the full set positioning solution and the subset positioning solution:

[0114]

[0115] Substitute equation (53) and equation (56) into equation (59), Δ i can be further expressed as:

[0116]

[0117] The solution separation Δ i obeys a Gaussian distribution:

[0118]

[0119] In step two, the constructed fault detection test statistic is:

[0120] Under the fault-free condition, the fault vector f is a zero vector, and Δ i obeys a Gaussian distribution with mean 0; however, when there is a fault, Δ i obeys a Gaussian distribution with non-zero mean. Therefore, Δ i is taken as the test statistic, and the satellite fault detection can be achieved by comparing it with a detection threshold;

[0121]

[0122] where D denotes the detection of a fault, and ND denotes the non-detection of a fault; q i denotes the fault detection test statistic, T i denotes the corresponding detection threshold; i = 0, 1, …, h is a set of mutually exclusive and exhaustive fault hypotheses, and h denotes the number of fault hypotheses;

[0123] In step two, the constructed fault exclusion test statistic is:

[0124] When a satellite fault is detected, fault exclusion needs to be performed to ensure the continuity of the navigation system. Select all detected fault subsets ‘i’ of the satellite as exclusion candidates, denoted as S i . Take S iEach satellite subset is taken as a new full set of visible satellites, and a multi-hypothesis resolution based fault detection test statistic is constructed to perform the second round of fault detection. If no fault is detected in a subset, it means that the fault is successfully eliminated;

[0125]

[0126] where E represents the fault elimination, NE represents the fault non-elimination; q i,j represents the fault elimination test statistic, T i,j represents the corresponding elimination threshold, j = 0, 1, …, h and j ≠ i.

[0127] In the third step, only in the case of fault detection, the integrity risk equation and the continuity risk equation are:

[0128] The integrity risk is equivalent to the probability of hazardous misleading information (HMI), which refers to the probability that the true position exceeds the maximum allowable position error boundary and is not detected. Therefore, the integrity risk equation is expressed as:

[0129]

[0130] where ε0 represents the horizontal positioning error or the vertical positioning error; corresponds to the horizontal alert limit or the vertical alert limit; P i represents the prior probability of the occurrence of the ith fault hypothesis;

[0131] The continuity risk is equivalent to the probability of loss of continuity (LOC), which refers to the probability of all events that cause the interruption of the task, including the detection of faults in the case of no fault and fault, as well as other events that cause loss of continuity. Therefore, the continuity risk equation is expressed as:

[0132]

[0133] In the third step, the integrity risk equation and the continuity risk equation are:

[0134] In the case of considering fault detection and elimination, a complete description of the integrity risk needs to consider all events that produce hazardous misleading information, including hazardous misleading information existing in the full set of positioning estimation solutions and hazardous misleading information existing in the subset of positioning estimation solutions after eliminating the fault satellite under a certain fault hypothesis j. Therefore, the integrity risk equation considering fault detection and elimination is expressed as:

[0135]

[0136] where, corresponding to the integrity risk generated by executing the fault elimination algorithm;

[0137] At the same time, continuity risk is used to describe the situation of task interruption when a fault is detected but cannot be eliminated. Therefore, the continuity risk equation under the condition of fault detection and elimination is expressed as:

[0138]

[0139] Compared with equation (65), equation (67) reduces the continuity risk by performing a fault elimination algorithm. However, the sacrifice for this is represented as , i.e. the integrity risk is increased;

[0140] The method of introducing a protection level to evaluate the integrity risk in step four is:

[0141] During the evaluation of the integrity risk, it is challenging to directly use equation (66) due to the relatively complex and computationally intensive joint probability distribution of the positioning error. Therefore, first, the upper bound of P HMI,FDE is derived:

[0142]

[0143] Under certain fault assumptions, the probability distribution of each term on the right side of equation (68) is known, expressed as:

[0144]

[0145] In the formula, Q(*) = 1 - Φ(*), Φ(*) represents the standard normal cumulative distribution function;

[0146] To meet the integrity requirement, the commonly used method is to calculate a position error boundary called the protection level (PL) so that the actual position has a large enough probability of falling within the error boundary. The protection level can be obtained by replacing l in equation (69) with PL and calculating P HMI,FDE to be within the integrity requirement;

[0147] P HMI,FDE ≤ A + B + C + D + E ≤ I req,x -P NM

[0148]

[0149]

[0150] In the formula, I req,x represents the integrity requirement allocated to the state x direction; P NM represents the prior probability of a fault that rarely occurs and does not need to be monitored;

[0151] As can be seen from equation (70), given the alarm limit l and the true navigation condition, the protection level is only related to B and E, i.e. only related to the detection threshold T i and the exclusion threshold T j,i . It is worth noting that B and E are monotone increasing functions of T i and T j,i , respectively;

[0152] In step five, in the process of allocating the continuity risk, the upper bound of P LOC,FDE is first derived:

[0153]

[0154] In the equation, the second term indicates that, in addition to other continuity losses P other , the continuity loss caused by fault detection and exclusion under each fault assumption constitutes a large part of the continuity requirement. Therefore, C req -P other can be equally allocated to each fault assumption:

[0155]

[0156] In fact, by introducing the parameter λ, C req,i can be further allocated to the fault detection and fault exclusion tests:

[0157] P{|q i |>T i | 0}P0≤C req,i,d = λ · C req,i (73)

[0158]

[0159] Then, according to equations (73) and (74), the detection threshold and the exclusion threshold are determined as:

[0160]

[0161] In the equation, Q -1 (*) represents the inverse tail probability distribution function of the two-tailed standard normal distribution;

[0162] According to equation (75), the detection threshold T i is a monotone decreasing function of λ, while the exclusion threshold T j,i is a monotone increasing function of λ. Recall the relationship between the protection level and T i and T j,i discussed in step four, it is further concluded that there must be a λ that makes the protection level minimum;

[0163] In step six, the method for determining the specific value of the key parameter λ is as follows:

[0164] To ensure greater accuracy of the determined parameters, the integrity risk was calculated for different values ​​of λ according to equation (69), where λ increases from 0 to 1 in increments of 0.01. Then, the λ corresponding to the minimum integrity risk was found, and the above two steps were repeated in each epoch. The final statistical results are shown in [reference needed]. Figure 1 . Figure 1 The left and right vertical axes can be viewed as the probability density and cumulative probability density of the key parameters, respectively.

[0165] Depend on Figure 1 It can be seen that the minimum integrity risk obtained at different epochs may correspond to different parameters. However, the range of these parameters will not exceed 0.46. Furthermore, from... Figure 1 It can be seen that 93.65% of the parameters fall within the range of 0.3. Therefore, it can be considered that the integrity risk is approximately minimized when the parameter is in the range of 0 to 0.3 within a 24-hour period. To further understand the impact of parameters within the range of 0.3 on integrity risk, the relationship between integrity risk and parameters was studied with parameter values ​​of 0.05, 0.15, and 0.25, as shown in [reference needed]. Figure 2 ;

[0166] In step six, based on the determined key parameter λ, the protection level is calculated using the bisection method according to equation (70) under the conditions of considering only fault detection and simultaneously considering fault detection and elimination. (See...) Figure 3 The availability of the ARAIM method is determined by comparing the obtained protection level with the alarm limit, as shown in Table 1.

[0167] Table 1 Availability of different methods

[0168]

[0169] Depend on Figure 3 As shown in Table 1, there is a difference in availability between the ARAIM method that only considers fault detection and the ARAIM method that considers both fault detection and troubleshooting. While the ARAIM method that only considers fault detection can provide the highest availability of 96.26%, it is susceptible to continuity risks, leading to unacceptable service interruption durations. Conversely, considering both fault detection and troubleshooting, the ARAIM method with a parameter of 0.5, although meeting the continuity requirement, only provides 83.48% availability. In contrast, the ARAIM algorithm based on the parameter of 0.15 determined in this invention is a better method, providing a higher availability of 88.70% and continuous navigation service.

[0170] The above specific embodiment description further details the purpose, technical solutions and beneficial effects of the application. It should be understood that the above description does not limit the patent protection scope of the application, and any equivalent transformation, modification or application in other related technical fields based on the content of the application description and drawings are also included in the patent protection scope of the application.

Claims

1. An advanced receiver autonomous integrity monitoring method considering fault detection and exclusion, characterized in that The method comprises the following steps: Step one, obtaining a positioning estimation solution of a linearized pseudo-range observation equation and an error of the positioning estimation solution; Step two, constructing a fault detection test statistic and a fault exclusion test statistic based on the positioning estimation solution and the error of the positioning estimation solution obtained in step one; Step three, establishing integrity risk equations and continuity risk equations according to the fault detection test statistic constructed in step two in the case of only considering fault detection, and establishing integrity risk equations and continuity risk equations according to the fault detection test statistic and the fault exclusion test statistic constructed in step two in the case of considering both fault detection and exclusion; Step four, performing integrity risk evaluation on the integrity risk equations established in step three by using a protection level; Step five, performing continuity risk allocation by introducing a key parameter λ to the continuity risk equations established in step three, and calculating fault detection thresholds and exclusion thresholds; Step six, determining the specific value of the key parameter λ in step five by a numerical method, and determining the protection level in step four according to the fault detection thresholds and the exclusion thresholds calculated in step five, to complete the advanced receiver autonomous integrity monitoring in the case of considering both fault detection and exclusion; In step four, the method for introducing the protection level to perform integrity risk evaluation is: In the process of assessing the integrity risk, first derive an upper bound on P HMI,FDE . The probability distribution of each term on the right side of formula (18) is known, and is expressed as: In the formula, Q(*) = 1-Φ(*), and Φ(*) represents a standard normal cumulative distribution function; The protection level is calculated by replacing I in formula (19) with PL and P HMI,FDE are calculated within the integrity requirements P HMI,FDE ≤ A + B + C + D + E ≤ I req,x -P NM where I req,x denotes the integrity requirement assigned to the direction of state x; P NM denotes the prior probability of a failure that occurs very rarely and does not need to be monitored The upper bound of P LOC,FDE is derived first in the process of continuous risk allocation. By introducing the parameter λ, C req,i can be further distributed to fault detection and troubleshooting tests: P{|q i |>T i |0}P0≤C req,i,d =λ·C req,i (23) Then, the detection thresholds and the exclusion thresholds are determined according to formulas (23) and (24): wherein Q -1 (*) denotes the two-tailed standard normal distribution function of the upper tail probability In step six, the method for determining the specific value of the key parameter λ is: The integrity risk under the condition of a given different λ is calculated according to formula (19), wherein λ increases from 0 to 1 at a step of 0.01, then the λ corresponding to the minimum integrity risk is found, and the above two steps are repeatedly performed in each epoch, and finally the key parameter is determined to be 0.05 or 0.15 according to the statistical results; Based on the determined key parameter λ, the protection level in the case of only considering fault detection and in the case of considering both fault detection and exclusion is calculated by using a bisection method according to formula (20), and the obtained protection level is compared with an alarm limit value.

2. The advanced receiver autonomous integrity monitoring method considering fault detection and exclusion according to claim 1, characterized in that: In step one, the linearized pseudo-range observation equation is: The linearized pseudo-range observation equation with n measurement values and m state values is expressed in the East-North-Up coordinate system as: z = Hx + v + f (1) where z is an n x 1 dimensional pseudo-range bias vector from the true position to the linearization point; H is an n x m dimensional geometry matrix from the positioning domain to the measurement domain; x is an m x 1 dimensional state vector containing the three dimensional position of the receiver and the clock parameter; v = [v1, v2,..., vn]Tis an n x 1 dimensional Gaussian observation noise vector with mean 0 and covariance matrix Σ; and b = [b1, b2,..., bn]Tis an n x 1 dimensional bias vector with mean 0 and covariance matrix R. n ] T is an n x 1 dimensional Gaussian observation noise vector with mean 0 and covariance matrix Σ. f is an n × 1-dimensional fault vector.

3. The advanced receiver autonomous integrity monitoring method considering fault detection and exclusion according to claim 1 or 2, characterized in that: In step one, the positioning estimation solution of the state vector x in the linearized pseudo-range observation equation and the error of the positioning estimation solution are calculated by using a weighted least squares algorithm, and the positioning estimation solution includes a full set positioning estimation solution and a subset positioning estimation solution; Let α be an m × 1-dimensional vector, which is used to extract a state of interest 'x' from the state vector full set x: A weighted least squares solution for the state x is obtained using all visible satellites and its positioning error ε0is In the formulae, W =∑ -1 is a weighting matrix, and the population positioning error e0obeys a Gaussian distribution: Define matrix W i With the same dimension as W, the diagonal elements corresponding to the row where the failed satellite is located are 0, and the weighted least squares solution of state x is obtained based on the subset of visible satellites And its positioning error ε i : In the formulae, subset positioning error ε i Subjection to Gaussian distribution: The solution separation definition is the difference between the full set positioning estimation solution and the subset positioning estimation solution: Substituting formula (3) and formula (6) into formula (9), Δ i can be further expressed as: Deconvolution Δ i Subjection to Gaussian distribution:

4. The method of claim 3, wherein the step two is characterized by: the constructed failure detection test statistics is: In the case of no fault, the fault vector f is a zero vector, and in step one Δ i obeys a Gaussian distribution with mean 0; when there is a fault, Δ i obeys a Gaussian distribution with non-zero mean, and Δ i satellite fault detection is achieved by comparing the test statistic Δ with a detection threshold. where D denotes detection of a fault, ND denotes no detection of a fault; q i denotes a fault detection test statistic, T i denotes a corresponding detection threshold; i = 0, 1, ···, h is a set of mutually exclusive and exhaustive fault hypotheses, h denotes the number of fault hypotheses.

5. The method of claim 4, wherein the step two is characterized by: the constructed failure exclusion test statistics is: When a satellite fault is detected, a subset 'i' of all detected faulty satellites is selected as exclusion candidates, denoted as S. i , with S i Each satellite subset is used as a new complete set of visible satellites. The fault detection test statistic is constructed by separating multiple hypothesis solutions to conduct a second round of fault detection. When no fault is detected in a certain subset, it means that the fault has been successfully eliminated. where E denotes an exclusion of a fault and NE denotes a non-exclusion of a fault; q i,j denotes a fault exclusion check statistic, T i,j denotes a corresponding exclusion threshold, j = 0, 1, ···, h and j ≠ i.

6. The method of claim 1 or 5, wherein the step three is characterized by: the integrity risk equation and the continuity risk equation are: the integrity risk equation is: In the formula, ε0 represents a horizontal positioning error or a vertical positioning error; corresponds to l represents a horizontal alarm limit or a vertical alarm limit; P i represents a prior probability of occurrence of the ith fault hypothesis; the continuity risk equation is:

7. The method of claim 6, wherein the step three is characterized by: the integrity risk equation is: In the formula, corresponding to the integrity risk resulting from executing the troubleshooting algorithm; the continuity risk equation is:

Images