阅读说明：本技术 一种基于冗余的惯导自检测系统 (Inertial navigation self-detection system based on redundancy ) 是由 邹慧 李明富 刘敏 姬海燕 王建岭 于 2019-09-09 设计创作，主要内容包括：本发明公开了一种基于冗余的惯导自检测系统,该系统包括传感器冗余、传感器数据检测、惯性算法冗余和基于残差平方和的投票选择机制四个步骤：传感器冗余选用三款不同类型的惯性传感器互为冗余；传感器数据检测又分为自检与互检两个过程；惯导算法冗余是在使用同组有效传感器测量值的情况下,采用两种或以上惯导解算算法实现惯导解算的过程；基于残差平方和的投票选择机制首先对两套惯导算法解算变量分别求取差值,并计算差值平方和。本发明的方式既能解决硬件损坏导致的惯导解算异常,也能通过多冗余算法获取多组解算结果,再通过投票选择机制判定飞控的健康状态,极大程度的提高了惯性导航系统的稳定性。(the invention discloses a redundancy-based inertial navigation self-detection system, which comprises four steps of sensor redundancy, sensor data detection, inertial algorithm redundancy and a voting selection mechanism based on residual sum of squares: the sensor redundancy selects three different types of inertial sensors which are redundant with each other; the sensor data detection is divided into two processes of self-detection and mutual detection; the redundancy of the inertial navigation algorithm is a process of realizing inertial navigation solution by adopting two or more inertial navigation solution algorithms under the condition of using the measurement values of the same group of effective sensors; the voting selection mechanism based on the residual sum of squares firstly solves the difference values of the two sets of inertial navigation algorithm resolving variables respectively and calculates the sum of the squares of the difference values. The method can solve the problem of inertial navigation solution abnormity caused by hardware damage, can also obtain a plurality of groups of solution results through a multi-redundancy algorithm, and judges the health state of flight control through a voting selection mechanism, thereby greatly improving the stability of the inertial navigation system.)

1. the utility model provides a system of inertial navigation self-test based on redundancy which characterized in that: the system comprises four steps of sensor redundancy, sensor data detection, inertial algorithm redundancy and a voting selection mechanism based on residual sum of squares, wherein:

S1, sensor redundancy: three different types of inertial sensors are selected as redundancy, and the different types of sensors have different characteristics, so that the sensor abnormality caused by the same sensor characteristic is avoided;

S2, sensor data detection: the detection is divided into two processes of self-detection and mutual detection, wherein the self-detection comprises the detection of effective output of a sensor and the detection of a measurement threshold; mutual detection is detection among a plurality of sensors, whether the data of the main sensor is normal or not is judged in a minority-compliant principle on the basis of a voting selection mode, and if the data of the main sensor is abnormal, redundant sensor data are used;

s3, redundancy of inertial navigation algorithm: the algorithm redundancy mechanism is a process of realizing inertial navigation resolving by adopting two or more inertial navigation resolving algorithms under the condition of using the measured values of the same group of effective sensors, and the adoption of the method is used for avoiding algorithm abnormity caused by resolving algorithm divergence or other reasons and greatly improving the stability of the inertial navigation system;

s4, voting selection mechanism based on sum of squared residuals: the voting selection mechanism based on the residual error is applied to sensor redundancy selection and algorithm redundancy, wherein the algorithm redundancy firstly respectively calculates difference values of two sets of inertial navigation algorithm resolving variables and calculates the sum of squares of the difference values, two sets of output data are close to each other under the normal condition of the two sets of independent resolving, the sum of squares of the residual error is small, two types of data are effective under the condition of being lower than a certain threshold value, and when the sum of squares of the residual error is large, the situation that at least one set of resolving data is abnormal currently exists is shown.

2. the redundancy-based inertial navigation self-detection system according to claim 1, wherein: the algorithm in S3 adopts EKF and UKF inertial navigation calculation algorithms respectively, mutual interactive supervision is relatively independent, iterative process variables are monitored in the calculation process and used as algorithm process monitoring, and the inertial navigation attitude and position data are independently calculated by the two algorithms so as to prepare for self-checking and selecting normal calculation data.

3. the redundancy-based inertial navigation self-detection system according to claim 1, wherein: the flow simplification process in S4 is as follows:

EKF resolving array: EKFOUT [9 ];

UKF resolves the array: UKFOUT [9 ];

Then the sum of the squared residuals between the two arrays is:

SUM＝EKFOUT[0]^2+EKFOUT[1]^2+…+EKFOUT[8]^2；

when the threshold is set, when SUM < a1, the two groups of data are both valid;

when the threshold is set, when SUM > a1, the following operations are performed:

calculating the three-axis attitude increment of the front and the back twice in a pure gyroscope integration mode: delta [3 ];

Deltangle[0]＝gyro[0]*dt；

Deltangle[1]＝gyro[0]*dt；

Deltangle[2]＝gyro[2]*dt；

calculating the angle variation of two times before and after under two calculation methods by recording the previous calculation data, and calculating the difference value of the three-axis postures of two times before and after under the same EKF algorithm to calculate delta1[3 ];

Calculating delta2[3] by the difference of the front and back triaxial postures under the same EKF algorithm;

and finally, taking a group of data with smaller residual square sum as effective inertial navigation resolving data by a mode of comparing the residual square sum:

EKFDeltsum＝(Deltangle[0]-delta1[0])^2+(Deltangle[1]-delta1[1])^2+(Deltangle[2]-delta1[2])^2；

UKFDeltsum＝(Deltangle[0]-delta2[0])^2+(Deltangle[1]-delta2[1])^2+(Deltangle[2]-delta2[2])^2；

And judging the validity of the current resolving data by judging the sizes of EKFDeltsum and UKFDeltsum.

Technical Field

The invention belongs to the technical field of inertial navigation, and particularly relates to a self-detection system for an inertial navigation resolving system, which is used for sensor redundancy, inertial navigation algorithm redundancy and a voting selection mechanism based on residual sum of squares.

background

The inertial navigation is based on Newton's inertial principle, and uses inertial element (accelerometer) to measure the acceleration of vehicle itself, and obtains speed and position through integration and operation, thereby achieving the purpose of navigation positioning of vehicle.

Currently, inertial navigation can be divided into two main categories: platform type inertial navigation and strapdown inertial navigation. The main difference is that the former has a physical platform, the gyro accelerometer is placed on the platform stabilized by the gyro, the platform tracks the navigation coordinate system to realize speed and position resolution, and the attitude data is directly taken from the ring frame of the platform; in the strap-down inertial navigation, a gyroscope and an accelerometer are directly and fixedly connected to a carrier. The function of the inertial platform is completed by a computer, and the attitude data of the inertial platform is obtained by calculation. Inertial navigation has a fixed drift rate, which causes an error of object motion, so that a long-range weapon usually adopts instructions, a GPS, and the like to perform timing correction on inertial navigation to acquire continuous and accurate position parameters. For example, the intermediate distance air-fuel missile adopts strapdown inertial navigation + instruction correction, the JDAM adopts an autonomous satellite positioning/inertial navigation combination (GPS/INS), the battle axe also adopts a GPS/INS + terrain matching technology, and most of the launch vehicles adopt platform inertial navigation, etc.

The mechanism of the inertial navigation system has developed various modes such as flexible inertial navigation, optical fiber inertial navigation, laser inertial navigation, micro solid-state inertial instrument and the like, and the inertial navigation system is widely applied to various aspects of aviation, aerospace, navigation and land maneuvering according to different requirements on environment and precision. In principle, an inertial navigation system is generally composed of an inertial measurement unit, a computer, a control display, and the like. The inertial measurement unit includes an accelerometer and a gyroscope, which is also called an inertial navigation combination. 3 degree-of-freedom gyroscopes are used to measure 3 rotational movements of the aircraft; the 3 accelerometers are used to measure the acceleration of 3 translational movements of the aircraft. The computer calculates the speed and position data of the aircraft according to the measured acceleration signal, and controls the display to display various navigation parameters.

the inertial navigation system often causes inertial navigation solution abnormity due to hardware damage, cannot be effectively removed in time, is not easy to judge the health state of flight control, and causes relatively low stability of the inertial navigation system. Therefore, it is desirable to provide a redundancy-based inertial navigation self-detection system that can improve the stability of an inertial navigation system.

disclosure of Invention

The invention aims to provide a redundancy-based inertial navigation self-detection system, and aims to solve the problems that in the prior art, inertia resolving abnormity is caused by hardware damage, and cannot be effectively and timely removed, so that the flight control health state is not easy to judge.

In order to achieve the purpose, the invention adopts the following technical scheme:

A redundancy-based inertial navigation self-detection system comprises four steps of sensor redundancy, sensor data detection, inertial algorithm redundancy and a voting selection mechanism based on the sum of squares of residual errors, wherein:

S1, sensor redundancy: three different types of inertial sensors are selected as redundancy, and the different types of sensors have different characteristics, so that the sensor abnormality caused by the same sensor characteristic is avoided;

s2, sensor data detection: the detection is divided into two processes of self-detection and mutual detection, wherein the self-detection comprises the detection of effective output of a sensor and the detection of a measurement threshold; mutual detection is detection among a plurality of sensors, whether the data of the main sensor is normal or not is judged in a minority-compliant principle on the basis of a voting selection mode, and if the data of the main sensor is abnormal, redundant sensor data are used;

s3, redundancy of inertial navigation algorithm: the algorithm redundancy mechanism is a process of realizing inertial navigation resolving by adopting two or more inertial navigation resolving algorithms under the condition of using the measured values of the same group of effective sensors, and the adoption of the method is used for avoiding algorithm abnormity caused by resolving algorithm divergence or other reasons and greatly improving the stability of the inertial navigation system;

S4, voting selection mechanism based on sum of squared residuals: the voting selection mechanism based on the residual error is applied to sensor redundancy selection and algorithm redundancy, wherein the algorithm redundancy firstly respectively calculates difference values of two sets of inertial navigation algorithm resolving variables and calculates the sum of squares of the difference values, two sets of output data are close to each other under the normal condition of the two sets of independent resolving, the sum of squares of the residual error is small, two types of data are effective under the condition of being lower than a certain threshold value, and when the sum of squares of the residual error is large, the situation that at least one set of resolving data is abnormal currently exists is shown.

preferably, the algorithm in S3 adopts two inertial navigation calculation algorithms, EKF and UKF, which are mutually independent and monitored, and the iterative process variables are monitored in the calculation process as the algorithm process monitoring, and the two algorithms independently calculate the inertial navigation attitude and position data to prepare for self-checking and selecting normal calculation data.

Preferably, the flow simplification process in S4 is as follows:

EKF resolving array: EKFOUT [9 ];

UKF resolves the array: UKFOUT [9 ];

then the sum of the squared residuals between the two arrays is:

SUM＝EKFOUT[0]^2+EKFOUT[1]^2+…+EKFOUT[8]^2；

when SUM < a1 (threshold set) then both sets of data are valid;

When SUM > a1 (threshold is set), the following operations are performed again:

calculating the three-axis attitude increment of the front and the back twice in a pure gyroscope integration mode: delta [3 ];

Deltangle[0]＝gyro[0]*dt；

Deltangle[1]＝gyro[0]*dt；

Deltangle[2]＝gyro[2]*dt；

Calculating the angle variation of two times before and after under two calculation methods by recording the previous calculation data, and calculating the difference value of the three-axis postures of two times before and after under the same EKF algorithm to calculate delta1[3 ];

calculating delta2[3] by the difference of the front and back triaxial postures under the same EKF algorithm;

and finally, taking a group of data with smaller residual square sum as effective inertial navigation resolving data by a mode of comparing the residual square sum:

EKFDeltsum＝(Deltangle[0]-delta1[0])^2+(Deltangle[1]-delta1[1])^2+(Deltangle[2]-delta1[2])^2；

UKFDeltsum＝(Deltangle[0]-delta2[0])^2+(Deltangle[1]-delta2[1])^2+(Deltangle[2]-delta2[2])^2；

And judging the validity of the current resolving data by judging the sizes of EKFDeltsum and UKFDeltsum.

the invention has the technical effects and advantages that: compared with the prior art, the redundancy-based inertial navigation self-detection system provided by the invention can solve the problem of inertial navigation solution abnormity caused by hardware damage, can also obtain multiple groups of solution results through a multi-redundancy algorithm, and judges the health state of flight control through a voting selection mechanism, thereby greatly improving the stability of the inertial navigation system.

drawings

FIG. 1 is a flow chart of the present invention for self-test and cross-test of sensor data;

Fig. 2 is a flow chart of voting selection by means of residual sum of squares according to the present invention.

Detailed Description

the technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. The specific embodiments described herein are merely illustrative of the invention and do not delimit the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The invention provides a redundancy-based inertial navigation self-detection system as shown in figures 1-2, which comprises four steps of sensor redundancy, sensor data detection, inertial algorithm redundancy and a voting selection mechanism based on residual sum of squares, wherein:

s1, sensor redundancy: three different types of inertial sensors are selected as redundancy, and the different types of sensors have different characteristics, so that the sensor abnormality caused by the same sensor characteristic is avoided; the three types of inertial sensors are respectively Sensor1, Sensor2 and Sensor3, and the three groups of inertial sensors are respectively provided with a self-checking device for detecting whether self-checking data and states are abnormal or not;

S2, sensor data detection: the detection is divided into two processes of self-detection and mutual detection, wherein the self-detection comprises the detection of effective output of a sensor and the detection of a measurement threshold; mutual detection is detection among a plurality of sensors, whether the data of the main sensor is normal or not is judged in a minority-compliant principle on the basis of a voting selection mode, and if the data of the main sensor is abnormal, redundant sensor data are used; the self-detection respectively comprises Sensor1 self-detection, Sensor2 self-detection and Sensor3 self-detection, three groups of self-detection data are judged by an abnormality processing judgment program, if the three groups of self-detection data are abnormal, the detection data result is abandoned, and if the three groups of self-detection data are not abnormal, the result is output; the mutual detection is that the data output of the effective Sensor is obtained by using the principle of minority-compliant majority through the voting selection mechanism among three groups of data, namely the non-abnormal output result of the self-detection of the Sensor1, the non-abnormal output result of the self-detection of the Sensor2 and the non-abnormal output result of the self-detection of the Sensor3 (as shown in fig. 1 in detail).

s3, redundancy of inertial navigation algorithm: the algorithm redundancy mechanism is a process of realizing inertial navigation resolving by adopting two or more inertial navigation resolving algorithms under the condition of using the measured values of the same group of effective sensors, and the adoption of the method is used for avoiding algorithm abnormity caused by resolving algorithm divergence or other reasons and greatly improving the stability of the inertial navigation system; the algorithm respectively adopts two inertial navigation calculation algorithms (shown in figure 2) of EKF and UKF, mutual interactive supervision is relatively independent, iterative process variables are monitored in the calculation process and used as algorithm process monitoring, and the two algorithms independently calculate inertial navigation attitude and position data so as to select normal calculation data for self-checking; the EKF and UKF algorithm results are subjected to algorithm self-monitoring flow, normal calculation data are output, and abnormal data are discarded after being distinguished;

S4, voting selection mechanism based on sum of squared residuals: the voting selection mechanism based on the residual error is applied to both sensor redundancy selection and algorithm redundancy, wherein the algorithm redundancy firstly respectively calculates difference values of two sets of inertial navigation algorithm calculation variables and calculates the sum of squares of the difference values (as shown in fig. 2), under the normal condition of the two sets of independent calculation, two sets of output data are close to each other, the sum of squares of the residual errors is small, two sets of data are both effective under the condition of being lower than a certain threshold value, and when the sum of the squares of the residual errors is large, the situation that at least one set of calculation data is abnormal currently exists is shown.

The specific simplified process is as follows:

EKF resolving array: EKFOUT [9 ];

UKF resolves the array: UKFOUT [9 ];

then the sum of the squared residuals between the two arrays is:

SUM＝EKFOUT[0]^2+EKFOUT[1]^2+…+EKFOUT[8]^2；

when SUM < a1 (threshold set, where a1 is the threshold), then both sets of data are valid;

when SUM > a1 (when a threshold is set, where a1 is the threshold), the following operations are performed:

calculating the three-axis attitude increment of the front and the back twice in a pure gyroscope integration mode: delta [3 ];

Deltangle[0]＝gyro[0]*dt；

Deltangle[1]＝gyro[0]*dt；

Deltangle[2]＝gyro[2]*dt；

calculating the angle variation of two times before and after under two calculation methods by recording the previous calculation data, and calculating the difference value of the three-axis postures of two times before and after under the same EKF algorithm to calculate delta1[3 ];

calculating delta2[3] by the difference of the front and back triaxial postures under the same EKF algorithm;

and finally, taking a group of data with smaller residual square sum as effective inertial navigation resolving data by a mode of comparing the residual square sum:

EKFDeltsum＝(Deltangle[0]-delta1[0])^2+(Deltangle[1]-delta1[1])^2+(Deltangle[2]-delta1[2])^2；

UKFDeltsum＝(Deltangle[0]-delta2[0])^2+(Deltangle[1]-delta2[1])^2+(Deltangle[2]-delta2[2])^2；

And judging the validity of the current resolving data by judging the sizes of EKFDeltsum and UKFDeltsum.

if EKFDeltsum is less than UKFDeltsum, the EKFDeltsum is effective data resolving;

if EKFDeltsum is more than UKFDeltsum, the UKFDeltsum is effective resolving data.

In summary, the following steps: the method can solve the problem of inertial navigation solution abnormity caused by hardware damage, can also obtain a plurality of groups of solution results through a multi-redundancy algorithm, and judges the health state of flight control through a voting selection mechanism, thereby greatly improving the stability of the inertial navigation system.

finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications may be made to the embodiments or portions thereof without departing from the spirit and scope of the invention.