CN110296701A - Inertia and satellite combined guidance system gradation type failure recall fault-tolerance approach - Google Patents

Abstract

The invention provides a gradual fault backtracking fault tolerance method of an inertial and satellite integrated navigation system. Carry out the initial alignment of the inertial navigation system; collect and slide the sensor data; use the state chi-square and residual chi-square hybrid detection method to detect and identify the gradual fault of the satellite navigation position information, when the satellite navigation has no fault, switch to inertial/ In the satellite integrated navigation mode, the satellite navigation output information is fused; when the satellite navigation is in a gradual failure, the inertial navigation backtracking algorithm and the Kalman backtracking algorithm are performed to obtain the attitude matrix, Speed and position; go back to the satellite navigation fault point, perform pure inertial navigation retrospective calculation, recurse to the kth sampling time, and output the attitude, speed and position information after isolating the historical fault information of the satellite. The invention can effectively judge the gradual failure of satellite navigation, and perform fault-tolerant processing on the failure by the inertial navigation and Kalman backtracking algorithm, so as to avoid the pollution of historical failure information.

Description

Gradual fault backtracking fault tolerance method for inertial and satellite integrated navigation system

technical field

The invention relates to a fault tolerance method of an inertial integrated navigation system, in particular to a gradual fault tolerance method of an inertial/satellite integrated navigation system.

Background technique

The inertial integrated navigation system integrates various sensor information, so the accuracy of the sensor information determines the system accuracy. The main task of the fault tolerance technology is to quickly isolate the faulty device when the external reference information fails, but because it is a gradual fault, the system will be polluted by historical fault information in a short time. When the fault occurs, the stored gyroscope and accelerometer information is used to perform the inertial navigation and Kalman retrospective algorithms, and after the historical fault information is eliminated, the pure inertial navigation rapid retrospective processing is carried out, and the historical gradual fault information of satellite navigation is effectively processed. Fault-tolerant, improving the accuracy and reliability of navigation results.

Among the published articles, such as "Research and Implementation of Fault Detection Algorithm for SINS/BD Tight Integrated Navigation System" by Yin Dequan, Xiong Zhi, Yang Jinghua and Min Yanling in "Navigation and Control", Volume 17, Issue 4, 2018 "In the article, an improved chi-square fault detection algorithm is mentioned, which uses the dynamic change characteristics of innovative information to identify gradual faults, and isolates the faulty satellites in real time, thereby effectively reducing the impact of fault information on inertial navigation accuracy. According to its simulation results, it can be concluded that this method can quickly monitor faulty satellites, and quickly avoid the pollution of fault information by setting different warning thresholds, but it can only isolate gradual faults in real time, and cannot further improve historical fault information. processing, the influence of historical gradual fault information on the system cannot be eliminated.

Also Zhang Huaqiang, Li Dongxing and Zhang Guoqiang put forward a hybrid χ ² detection method in the article "Application of Hybrid χ ² Detection Method in Integrated Navigation System Fault Detection" published in "Chinese Journal of Inertial Technology". The χ ² detection method uses parallel operation to detect gradual faults of sensor information, and the fault diagnosis result of the integrated navigation system is jointly determined by the results of the parallel operation, which can effectively reduce the false alarm rate, accurately isolate the gradual fault information, and effectively improve the fault diagnosis of the integrated navigation system. As a result, system reliability is improved. From the author's simulation results, it can be seen that the integrated navigation system in fault-tolerant processing can effectively detect the gradual failure of satellite navigation information due to the introduction of fault information after 100-110 s after the failure of satellite navigation information simulation. After the fault, no further fault-tolerant processing is performed on the previously fused historical gradual fault information, and the system is still polluted by the fault information within a certain period of time, and the steady-state impact of the historical gradual fault on the subsequent system cannot be eliminated.

The above published articles have all described and explored the gradual fault tolerance of the integrated navigation system, but they have not achieved the further fault tolerance processing of the historical fault information after the gradual fault is detected. The method of eliminating historical faults and recalculating the navigation information in the fault period has innovative and practical engineering value.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide an inertial and satellite integrated navigation system gradual fault retrospective fault tolerance method that can effectively judge satellite navigation gradual faults, perform fault-tolerant processing on the faults, and avoid historical fault information contaminating the system.

The object of the present invention is achieved in this way:

Step 1, the inertial integrated navigation system, binding the initial navigation parameters, and performing the initial alignment of the inertial navigation system;

Step 2, collect sensor data, and slide and store the sensor data in the [t _kM ,t _k ], M=h/T time period, mainly including gyroscope information ω ^b , accelerometer information f ^b and satellite navigation latitude, longitude, altitude information Among them, t _k is the kth sampling time, M is the number of backtracking, h is the backtracking time, T _s is the sampling period, and b represents the carrier coordinate system;

Step 3, according to the satellite navigation position information, use the state chi-square and residual chi-square mixed detection method to perform data analysis and processing, to detect and identify the gradual fault, when the satellite navigation has no gradual fault, go to step 4; When the navigation is in a gradual failure, go to step 5;

Step 4, the system switches into the inertial/satellite integrated navigation mode, uses the Kalman filter algorithm to update the time and measurement, and updates the satellite navigation information at the kth sampling time. Fusion is performed to output the corrected body position, velocity and attitude information;

Step 5, the attitude matrix from the carrier coordinate system b to the navigation coordinate system n at the kth sampling time of the inertial navigation system speed information and location information Use the gyroscope information stored at the k-th sampling time in sliding and accelerometer information Carry out the inertial navigation backtracking algorithm to obtain the body attitude matrix at the k-1 sampling time speed and location

Step 6, perform the Kalman filter backtracking algorithm, and use the position information calculated by the inertial navigation at the k-1 sampling time Utilize the position information of the satellite navigation stored at the k-1th sampling time by sliding Obtain the Kalman backtracking measurement information at the k-1 sampling time:

Step 7, estimate the state of the Kalman filter at the kth sampling time and the mean square error of the state estimate P _k , perform time retrospective update, and recursively predict the state at the k-1th sampling time by one step and state one-step prediction mean square error P _k-1/k :

in, Φ _k/k-1 is the state one-step transition matrix of the Kalman filter from the k-1th sampling time to the kth sampling time; Γ _k-1 is the system noise driving matrix at the k-1 sampling time; Q _k-1 is the system noise matrix at the k-1 sampling time;

Step 8: Backward updating of the Kalman measurement is performed based on the retrospective measurement information and time retrospective update:

P _k-1 =(IK _k-1 H _k-1 )P _k-1/k

Among them, K _k-1 is the filter gain matrix at the k-1 sampling time; H _k-1 is the measurement matrix at the k-1 sampling time; R _k-1 is the measurement noise matrix at the k-1 sampling time; I is the identity matrix;

Step 9, according to the error estimator obtained by the Kalman backtracking algorithm Feedback the position information corresponding to the k-1 sampling time of the inertial navigation backtracking solution, and obtain the attitude matrix without fusion of the satellite gradual fault information at the k-1 sampling time. speed and location

Step 10: Go back to the satellite navigation fault point, use the gyroscope and accelerometer information stored from the fault point to the kth sampling time in the [t _kM ,t _k ] time period to perform pure inertial navigation forward retrospective calculation, recursive At the kth sampling time, the body attitude, speed and position information after isolating the historical fault information of the satellite are output.

The present invention provides a new method suitable for inertial integrated navigation system, capable of effectively judging satellite navigation gradual faults, and performing inertial navigation and Kalman backtracking algorithms to perform fault-tolerant processing on faults and avoid historical fault information contaminating the system.

The advantages of the present invention are mainly reflected in:

Aiming at the problem of gradual faults in the inertial/satellite integrated navigation system, the present invention provides an inertial navigation and Kalman backtracking algorithm suitable for fault-tolerant processing of gradual faults, which can effectively eliminate historical fault information and effectively improve the precision and accuracy of the integrated navigation system. reliability.

Description of drawings

FIG. 1 is a flow chart of a fault-tolerance method for gradual fault backtracking applicable to an inertial/satellite integrated navigation system according to the present invention.

Detailed ways

The present invention will be described in more detail with examples below.

Step 1, start the inertial integrated navigation system, and bind the initial navigation parameters: the constant offset ε _x ,ε _y ,ε _z of the inertial measurement device gyroscope and the constant offset ▽ _x ,▽ _y ,▽ _z of the accelerometer, satellite Navigation position measurement error initial latitude Initial longitude λ ₀ , initial height h ₀ ; perform the initial alignment of the inertial navigation system;

Step 2: Collect sensor data, and store the sensor data in the time period of [t _kM ,t _k ], (M=h/T _s ), mainly including gyroscope information ω ^b , accelerometer information f ^b and satellite navigation latitude , longitude, altitude information Among them, t _k is the kth sampling time, M is the number of backtracking, h is the backtracking time, T _s is the sampling period, and b represents the carrier coordinate system;

Step 3, according to the satellite navigation position information, use the state chi-square and residual chi-square mixed detection method to perform data analysis and processing, to detect and identify the gradual fault, when the satellite navigation has no fault, go to step 4; when the satellite navigation is in In case of gradual failure, go to step 5;

Step 4, the system switches into the inertial/satellite integrated navigation mode, uses the Kalman filter algorithm to update the time and measurement, and outputs information for the satellite navigation at the kth sampling time. Fusion is performed to output the corrected body position, velocity and attitude information;

Step 5, the attitude matrix from the carrier coordinate system b to the navigation coordinate system n at the kth sampling time of the inertial navigation system speed information and location information Use the gyroscope information at the k-th sampling moment stored by sliding Accelerometer information Backtracking the inertial navigation attitude, get the attitude matrix at the k-1 sampling time:

in, × stands for The antisymmetric matrix of , is the projection of the rotational angular rate of the carrier coordinate system b relative to the navigation coordinate system n under the carrier coordinate system b at the kth sampling time; is the projection of the earth's rotation angular rate at the kth sampling time under the navigation coordinate system n; is the projection of the rotational angular velocity of the navigation coordinate system n relative to the earth coordinate system e at the k-th sampling time under the navigation coordinate system n; R _M is the main radius of curvature of the local earth meridian; R _N is the main radius of curvature of the local earth's meridian circle; is the projection of the airframe velocity in the navigation coordinate system n in the east, north and sky directions at the kth sampling time;

Step 6, get the attitude matrix at the k-1 sampling time Backtrack to calculate the body speed at the k-1 sampling time:

in, is the projection of the local gravitational acceleration under the navigation coordinate system n at the kth sampling time;

Step 7, from the speed information of the k-1 sampling time Backtrack to calculate the position of the body at the k-1 sampling time:

Step 8, carry out the Kalman filter backtracking algorithm, the position information calculated by the inertial navigation backtracking at the k-1th sampling time Utilize the position information of the satellite navigation stored at the k-1th sampling time by sliding Obtain the Kalman backtracking measurement information at the k-1 sampling time:

Step 9, estimate the state of the Kalman filter at the kth sampling time and the mean square error of the state estimate P _k , perform time retrospective update, and recursively predict the state at the k-1th sampling time by one step and state one-step prediction mean square error P _k-1/k :

in, Φ _k/k-1 is the state one-step transition matrix of the Kalman filter from the k-1th sampling time to the kth sampling time; Γ _k-1 is the system noise driving matrix at the k-1 sampling time; Q _k-1 is the system noise matrix at the k-1 sampling time, and Q _k-1 =Q _k =...=Q ₀ is a constant value matrix;

Step 10, perform the Kalman measurement retrospective update based on the retrospective measurement information and time retrospective update:

P _k-1 = (IK _k-1 H _k-1 )P _k-1/k (13)

Among them, K _k-1 is the filter gain matrix at the k-1 sampling time; H _k-1 is the measurement matrix at the k-1 sampling time; R _k-1 is the measurement noise matrix at the k-1 sampling time, and R _{k -1} =R _k =...=R ₀ is a constant value matrix; I is a unit matrix;

Step 11, according to the error estimator obtained by the Kalman backtracking algorithm Feedback the position information corresponding to the k-1 sampling time of the inertial navigation backtracking solution, and obtain the attitude matrix without fusion of the satellite gradual fault information at the k-1 sampling time. speed and location

Step 12: Alternately perform inertial navigation and Kalman backtracking algorithms to backtrack to the satellite navigation fault point; use the gyroscope and accelerometer information stored from the fault point to the kth sampling time in the [t _kM ,t _k ] time period to perform pure The inertial navigation forward retrospective solution, recursively to the kth sampling time, outputs the body attitude, speed and position information after isolating the historical fault information of the satellite, and completes the fault-tolerant processing of the gradual historical fault of the satellite navigation.

The described fault-tolerance method for gradual fault backtracking suitable for inertial/satellite integrated navigation systems adopts the hybrid chi-square detection method, which can quickly and effectively detect the gradual fault information of satellite navigation. The inertial navigation backtracking and Kalman backtracking algorithms are used to eliminate the The historical fault information in the system further improves the reliability of the system.

The method uses the sliding storage method to store the key data of the inertial navigation, which effectively saves the storage space, and uses the pure inertial navigation retrospective method, which can quickly output the navigation information of the airframe while avoiding the pollution of the historical fault information of the satellite. practicality of the project.

Claims (1)

1. A gradual-change fault backtracking fault-tolerant method for an inertia and satellite integrated navigation system is characterized by comprising the following steps:

step 1, an inertial integrated navigation system binds initial navigation parameters and performs initial alignment of the inertial navigation system;

step 2, collecting sensor data, and storing [ t ] in a sliding manner_k-M,t_k]The sensor data in the h/T time period mainly includes gyroscope information ω^bAccelerometer information f^bAnd satellite navigation latitude, longitude, altitude informationWherein, t_kIs the kth sampling time, M is the backtracking number, h is the backtracking time, T_sB represents a carrier coordinate system for a sampling period;

step 3, according to the satellite navigation position information, data analysis processing is carried out by using a state chi-square and residual chi-square mixed detection method, detection and identification are carried out on the gradual-change type fault, and when the satellite navigation has no gradual-change type fault, the step 4 is carried out; when the satellite navigation is in the gradual change type fault, performing step 5;

step 4, the system is switched into an inertia/satellite integrated navigation mode, time updating and measurement updating are carried out by using a Kalman filtering algorithm, and satellite navigation information of the kth sampling moment is subjected toFusing, and outputting the corrected body position, speed and posture information;

step 5, an attitude matrix from the k sampling moment carrier coordinate system b of the inertial navigation system to the navigation coordinate system nSpeed informationAnd location informationGyroscope information at kth sampling time stored by slidingAnd accelerometer informationPerforming inertial navigation backtracking algorithm to obtain an organism attitude matrix at the k-1 sampling momentSpeed of rotationAnd position

Step 6, performing a Kalman filter backtracking algorithm, and resolving position information by inertial navigation at the k-1 th sampling momentPosition information of satellite navigation using sliding stored k-1 th sampling timeObtaining Kalman backtracking measurement information at the k-1 sampling moment:

step 7, estimating the state of the Kalman filter at the kth sampling momentSum state estimation mean square error P_kPerforming time backtracking update, and recurrently predicting the state at the k-1 th sampling moment by one stepSum-state one-step prediction mean square error P_k-1/k：

Wherein,Φ_k/k-1a state one-step transition matrix of the Kalman filter from the k-1 sampling moment to the k sampling moment is obtained;Γ_k-1driving a matrix for the system noise at the k-1 th sampling moment; q_k-1A system noise matrix at the k-1 sampling moment;

and 8, performing Kalman measurement backtracking update by backtracking measurement information and time backtracking update:

P_k-1＝(I-K_k-1H_k-1)P_k-1/k

wherein, K_k-1Filtering a gain matrix at the k-1 th sampling moment; h_k-1A k-1 sampling moment measurement matrix is obtained; r_k-1Measuring a noise matrix at the k-1 sampling moment; i is an identity matrix;

step 9, obtaining error estimator according to Kalman backtracking algorithmFeeding back the position information corresponding to the k-1 sampling moment of inertial navigation backtracking calculation to obtain an attitude matrix which is not fused with the gradual change type fault information of the satellite at the k-1 sampling momentSpeed of rotationAnd position

Step 10, backtracking to the satellite navigation fault point, and storing the fault point in [ t ]_k-M,t_k]And (3) carrying out forward tracing resolving on the information of the gyroscope and the accelerometer from the fault point to the kth sampling time in the time period, recurrently obtaining the information of the gyroscope and the accelerometer to the kth sampling time, and outputting the body attitude, the speed and the position information after isolating the historical fault information of the satellite.