External field self-calibration method for double-shaft rotation inertial navigation system of helicopter

Abstract

The invention provides an outfield self-calibration method for a double-shaft rotating inertial navigation system of a helicopter, which comprises the following steps: controlling a double-shaft indexing mechanism of the double-shaft rotating inertial navigation system to execute an eight-position indexing scheme, carrying out navigation calculation to obtain a first speed error, and carrying out estimation by adopting a first Ka lman filter to obtain a first-stage parameter calibration result; a double-shaft indexing mechanism of the shaft rotation inertial navigation system periodically rotates according to a six-sequence continuous indexing scheme, navigation calculation is carried out, a first-stage parameter calibration result is adopted for parameter compensation, a second speed error is obtained, a second Ka lman filter is adopted for estimating an inner lever arm, and a second-stage parameter calibration result is obtained; the first-stage parameter calibration result and the second-stage parameter calibration result are bound to the double-shaft rotating inertial navigation system, so that the problem that a user needs to regularly return the inertial navigation system to a supplier for calibration is solved, the maintainability of a product is greatly improved, and the maintenance stop time of a helicopter is shortened.

Description

Technical Field

[0001] The present application belongs to the field of dual-axis rotating inertial navigation system self-calibration, and in particular to an outfield self-calibration method for a helicopter dual-axis rotating inertial navigation system. Background Art

[0002] As a key system for measuring the key motion parameters of helicopters and calculating the most basic navigation parameters such as helicopter attitude, speed, and position, the helicopter inertial navigation system is an important prerequisite for helicopters to perform flight missions and ensure flight safety. The dual-axis rotational inertial navigation system (RINS) is equipped with a dual-axis rotation mechanism outside the inertial measurement unit, which allows the inertial measurement unit to rotate according to the set rotation scheme. The symmetrical rotation motion can offset the sensor error to improve the navigation accuracy. The working principle block diagram is shown in Figure 1 .

[0003] According to the maintenance requirements of the inertial navigation system in the helicopter avionics system, after the component has been used for a given flight time or calendar time (whichever comes first), it must be calibrated separately by the supplier using a dedicated turntable. However, since the calibration work requires the inertial navigation system to be disassembled, sent to the supplier, calibrated, returned to the aircraft, installed, and powered on for inspection, the entire process takes from several weeks to several months, causing inconvenience to the actual use of helicopter users.

[0004] The parameters that need to be calibrated in the rotating inertial navigation system mainly include gyro constant drift, accelerometer constant bias, gyro scale factor error, accelerometer scale factor error, gyro installation error, accelerometer installation error, and inner lever arm. The field self-calibration of the above parameters has been achieved in the three-axis rotating inertial navigation system through parameter estimation based on Kalman filtering. However, in the two-axis rotating inertial navigation system, due to the limitation that the two-axis rotation mechanism can only provide two-axis rotation, the rotation scheme in the three-axis rotating inertial navigation system cannot be used to fully excite the parameters to be calibrated and make the Kalman filter state vector completely observable. It is necessary to redesign the Kalman filter and the two-axis rotation scheme.

[0005] In the research on the self-calibration method of dual-axis rotating inertial navigation system, most of them only calibrate some parameters other than the inner lever arm. In the self-calibration method considering all the parameters of the inner lever arm, in order to make the Kalman filter state vector completely observable, either the carrier is required to cooperate in maneuvering, or the number of calibration of the inner lever arm parameters is simplified, and only three axial components are calibrated.

[0006] The current helicopter dual-axis rotation inertial navigation system has a poor self-calibration effect and there is still much room for improvement. Summary of the invention

[0007] The present invention proposes an off-site self-calibration method for a helicopter dual-axis rotation inertial navigation system, which solves the pain point problem that users need to return the inertial navigation system to the supplier for calibration on a regular basis, greatly improves the maintainability of the product, and shortens the maintenance and downtime of the helicopter.

[0008] The present invention provides an outfield self-calibration method for a helicopter dual-axis rotation inertial navigation system, comprising:

[0009] Step 1: Perform a power-on check on the dual-axis rotation inertial navigation system, and align it if the power-on check result is normal;

[0010] Step 2, controlling the dual-axis indexing mechanism of the dual-axis rotation inertial navigation system to execute the eight-position indexing scheme, and performing navigation solution to obtain a first velocity error, using the first velocity error as the observation Z of the first Kalman filter, and using the first Kalman filter to estimate the gyro constant drift, the accelerometer constant bias, the gyro scale factor error, the accelerometer scale factor error, the gyro installation error, and the accelerometer installation error to obtain the first stage parameter calibration result;

[0011] Step 3, the dual-axis indexing mechanism of the axis rotation inertial navigation system rotates periodically according to the six-order continuous indexing scheme, and performs navigation solution and parameter compensation using the first-stage parameter calibration results to obtain a second velocity error, and uses the second velocity error as the observation Z of the second Kalman filter, and uses the second Kalman filter to estimate the inner lever arm to obtain the second-stage parameter calibration results;

[0012] Step 4: Bind the first-stage parameter calibration results and the second-stage parameter calibration results to the dual-axis rotation inertial navigation system.

[0013] Optional, eight-position indexing solutions include:

[0014] S11, placed in the initial position; in the initial position, x s The axis points east, y s The axis points north, z s The axis points to the sky;

[0015] S12, around y s The axis rotates 270° counterclockwise;

[0016] S13, around y s The axis rotates 180° clockwise;

[0017] S14, around z s The axis rotates 270° counterclockwise;

[0018] S15, around z s The axis rotates 180° clockwise;

[0019] S16, around y s The axis rotates 180° clockwise;

[0020] S17, go around z first s The axis rotates 45° clockwise; then rotates around y s The axis rotates 45° clockwise;

[0021] S18, around y s The axis rotates 180° clockwise;

[0022] x s Axis, y s Axis, z s The axes are the ideal sensitive axis directions of the x-axis gyro, y-axis gyro and z-axis gyro, and the outer rotation axis and inner rotation axis of the double-axis indexing mechanism are respectively s Axis and z s Axis coincidence; Ox s y s z s The origin O of the coordinate system is the rotation center of the double-axis indexing mechanism.

[0023] Optional, six-sequence sequential transposition schemes include:

[0024] S21, around y s Axis and z s The axes simultaneously perform uniform angular motion at a first preset angular acceleration rate and continue for a preset duration;

[0025] S22, around y s The axis makes a uniform angular motion at the first angular velocity; at the same time, around the z s The axis performs a uniform deceleration angular motion at a first preset angular acceleration rate and continues for a preset time period;

[0026] S23, around y s The axis performs a uniform deceleration angular motion at a first preset angular acceleration rate and continues for a preset time period;

[0027] S24, around z s The axis performs uniform angular motion at a first preset angular acceleration rate and continues for a preset time period;

[0028] S25, around y s The axis performs uniform acceleration angular motion at a first preset angular acceleration rate; at the same time, around the z s The axis performs uniform angular motion at a first angular rate and continues for a preset time.

[0029] S26, around y s Axis and z s The axes simultaneously perform uniform deceleration angular motion at a first preset angular acceleration rate and continue for a preset duration.

[0030] Optionally, the first preset angular acceleration

[0031] The first angular velocity is determined according to a first preset angular acceleration rate and a preset duration.

[0032] Optionally, the first Kalman filter includes a first state equation and a first observation equation;

[0033] The first state equation is

[0034] Where X is the first state vector,

[0035] φ n ,δV n They are the first attitude error and the first velocity error respectively;

[0036] The first attitude error and first velocity error equations of the inertial navigation system are:

[0037]

[0038] In the formula,

[0039] n, s, i, e in superscript and / or subscript represent navigation coordinate system, rotation coordinate system, geocentric inertial coordinate system, and earth coordinate system, respectively;

[0040] Represents the direction cosine matrix from the rotation coordinate system to the navigation coordinate system;

[0041] is the Earth's rotation angular velocity, is the angular velocity caused by the motion of the carrier;

[0042] f represents the specific force output of the accelerometer;

[0043] δω s is the gyro error, and its error model is:

[0044]

[0045] is the gyro constant drift;

[0046] S g =[s gx s gy s gz ] T is the gyro scale factor error;

[0047] is the angular velocity of the rotating coordinate system relative to the inertial coordinate system;

[0048] Μ g =[α xz α xy α yz α yx α zy α zx ] T is the gyro installation error;

[0049] v g is the gyro angular rate white noise;

[0050] δf s is the accelerometer error, and its error model is:

[0051]

[0052] is the constant bias of the accelerometer;

[0053] S a =[s ax s ay s az ] T is the accelerometer scale factor error;

[0054] Μ a =[β xz β xy β yz β yx β zy β zx ] T is the accelerometer installation error.

[0055] v a is the accelerometer comparative white noise;

[0056] F is the first system matrix, according to the attitude and velocity error equations:

[0057]

[0058] diag() means converting a 3×1 vector into a 3×3 matrix with vector elements on the diagonal;

[0059] fun() means converting a 3×1 vector into the following 3×6 matrix;

[0060]

[0061] W is the white noise of the gyroscope and accelerometer;

[0062]

[0063] The first observation equation is:

[0064] Z = HX + v;

[0065] Where Z is the first observation vector, Z = δV n ;

[0066] H is the first observation matrix, H = [0 3×3 Ι 3×3 0 3×24 ], v is the observation noise.

[0067] Optionally, the second Kalman filter includes a second state equation and a second observation equation; the second state equation is:

[0068]

[0069] in,

[0070] The second state vector

[0071] They are the second attitude error and the second velocity error respectively;

[0072]

[0073] According to the attitude and velocity error equations, the system matrix F2 can be obtained:

[0074]

[0075] The second observation equation is:

[0076] Z2=H2X2+v

[0077] Among them, Z2 is the observation vector,

[0078] H2 is the observation matrix, H2=[0 3×3 Ι 3×3 0 3×9 ];

[0079] r R =[(r X ) T (r Y ) T (r Z ) T ] T ;

[0080] r R is the inner lever arm, r X 、r Y 、r Z They represent the error vectors in three directions between the sensitive points of the three accelerometers and the rotation center point of the dual-axis indexing mechanism, rX =[r Xx r Xy r Xz ] T 、r Y =[r Yx r Yy r Yz ] T 、r Z =[r Zx r Zy r Zz ] T ;

[0081] δf=M vR r R ;

[0082]

[0083] ω is and Represents the angular velocity and angular acceleration of the rotating coordinate system relative to the inertial coordinate system, ω is =[ω x ω y ω z ] T ;

[0084] U X , U Y and U Z Represents the unit vector corresponding to each accelerometer's sensitive axis.

[0085] Optionally, before step 4, the method further includes:

[0086] Determine whether the difference between the first-stage parameter calibration result, the second-stage parameter calibration result and the historical parameter calibration result is within a preset range.

[0087] Optionally, before step 3, the method further includes:

[0088] Determine that the difference between the first-stage parameter calibration result and the corresponding historical parameter calibration result is within a preset range;

[0089] Correspondingly, before step 4, the method further includes:

[0090] Determine whether the difference between the second-stage parameter calibration result and the corresponding historical parameter calibration result is within a preset range.

[0091] Beneficial technical effects of this application:

[0092] The present invention proposes an off-site self-calibration method for a helicopter dual-axis rotating inertial navigation system. According to the characteristic that the inner lever arm is only excited when the indexing mechanism rotates, the parameters to be calibrated are divided into two categories, and the calibration is completed in two stages. The off-site self-calibration of 33 parameters of the helicopter dual-axis rotating inertial navigation system, including 9 inner lever arm parameters, can be realized, solving the pain point problem that users need to return the inertial navigation system to the supplier for regular calibration, greatly improving the maintainability of the product and shortening the maintenance and downtime of the helicopter. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 This is the block diagram of the working principle of the dual-axis RINS;

[0094] Figure 2 This is the flow chart of the self-calibration algorithm for the dual-axis rotation inertial navigation system;

[0095] Figure 3 Schematic diagram of the eight-position transposition scheme in the first stage of the dual-axis RINS self-calibration algorithm;

[0096] Figure 4 It is a schematic diagram of the inner lever arm;

[0097] Figure 5 Schematic diagram of the six-order transposition scheme in the second stage of the dual-axis RINS self-calibration algorithm. DETAILED DESCRIPTION

[0098] The specific contents of the technical solution provided by the present invention are introduced below in conjunction with the accompanying drawings.

[0099] See also Figure 1-5 The present invention provides a dual-axis RINS self-calibration method, which realizes the self-calibration of a total of 33 parameters including 9 inner lever arm parameters under the static base of the helicopter. The idea of ​​this method is to divide the parameters to be calibrated into two categories according to the characteristic that the inner lever arm is only excited when the transfer mechanism rotates, and complete the calibration in two stages. The first category is constant drift, accelerometer constant bias, gyro scale factor error, accelerometer scale factor error, gyro installation error, accelerometer installation error, totaling 24 parameters. The second category is 9 inner lever arm parameters. The dual-axis rotation inertial navigation system self-calibration algorithm of the present invention designs Kalman filter 2 and six-order continuous transfer scheme for the calibration of 9 inner lever arm parameters, and proves through the PWCS observability analysis method that under the six-order continuous transfer scheme, the 9 inner lever arm parameters are fully excited and the state vector of Kalman filter 2 is completely observable. .

[0100] The present invention proposes a self-calibration method, which mainly includes the following technical steps:

[0101] Step 1: Preparation: Power on the dual-axis rotating inertial navigation system and check that it is normal, and perform initial alignment.

[0102] Step 2: After the initial alignment is completed, execute the dual-axis rotation inertial navigation system self-calibration algorithm (see Figure 2 ).

[0103] Step 3: The technicians shall check the parameter results output by the dual-axis inertial navigation system self-calibration algorithm and bind them to the dual-axis rotation inertial navigation system. Note: If the difference between the parameter result value of this calibration and the parameter result value of previous calibrations exceeds the given value, it is still necessary to return to the supplier for calibration using a dedicated turntable.

[0104] Step 2 includes two stages:

[0105] Phase 1: Calibrate the first type of parameters. The dual-axis rotation inertial navigation system starts navigation solution. The dual-axis indexing mechanism remains stationary at position 1 specified in the eight-position indexing scheme for 5 minutes, and Kalman filter 1 starts state estimation. After 5 minutes, the rotation of the switching position is executed, and Kalman filter 1 skips observation update. After the rotation ends and the next position is stationary, Kalman filter 1 resumes observation update. By analogy, the eight-position indexing scheme is completed, the first phase ends, and the calibration results of the first type of parameters are obtained from the state vector.

[0106] Phase 2: Calibrate the second type of parameters. The dual-axis rotation inertial navigation system uses the calibration results of the first type of parameters to perform parameter compensation during the navigation solution process. The dual-axis indexing mechanism rotates periodically according to the six-order continuous indexing scheme, and Kalman filter 2 starts state estimation. After executing the six-order continuous indexing scheme 20 times, the second phase ends and the calibration values ​​of the second type of parameters are obtained from the state vector.

[0107] Specifically, step 2 may include: 2 stages.

[0108] In the first stage, the dual-axis rotation inertial navigation system starts navigation solution, and the dual-axis transfer mechanism executes the eight-position transfer scheme. The gyro constant drift, accelerometer constant bias, gyro scale factor error, accelerometer scale factor error, gyro installation error, and accelerometer installation error are estimated using the Kalman filter 1 with the speed error of the navigation solution as the observation to obtain the parameter calibration result.

[0109] In the second stage, the dual-axis rotation inertial navigation system uses the parameter calibration results of the first stage to perform parameter compensation in the navigation solution process. The dual-axis rotation mechanism rotates periodically according to the six-order continuous rotation scheme, and the inner lever arm is estimated using the Kalman filter 2 with the speed error of the navigation solution as the observation to obtain the parameter calibration result.

[0110] Exemplarily, the eight-position indexing scheme performed in the first stage refers to rotating the Ox through the dual-axis indexing mechanism. s y s z s The coordinate system is in 8 different positions in sequence. s y s z s The coordinate system represents a rotating coordinate system. The origin O of the coordinate system is the rotation center of the dual-axis indexing mechanism. s Axis, y s Axis, z s The axis is defined as the ideal sensitive axis direction of the x-axis gyro, y-axis gyro and z-axis gyro. The outer rotation axis and inner rotation axis of the dual-axis indexing mechanism are respectively s Axis and z s Axes coincide.

[0111] The eight-position transposition scheme is defined as shown in Table 1 (see Figure 3 ):

[0112] Table 1 Definition of the eight-position transposition scheme

[0113]

[0114] Exemplarily, in stage 1, the constructed Kalman filter 1 consists of a state equation and an observation equation.

[0115] The state equation is:

[0116]

[0117] in,

[0118] State Vector

[0119] φ n ,δV n They are attitude error and velocity error respectively;

[0120] The attitude and velocity error equations of the inertial navigation system are:

[0121]

[0122] In the formula,

[0123] The superscripts n, s, i, and e represent the navigation coordinate system, the rotating coordinate system, the geocentric inertial coordinate system, and the earth coordinate system, respectively;

[0124] Represents the direction cosine matrix from the rotation coordinate system to the navigation coordinate system;

[0125] is the Earth's rotation angular velocity, is the angular velocity caused by the motion of the carrier, is the angular velocity of the rotating coordinate system relative to the inertial coordinate system;

[0126] f represents the specific force output of the accelerometer;

[0127] δω s is the gyro error, and its error model is:

[0128]

[0129] is the gyro constant drift;

[0130] S g =[s gx s gy s gz ] T is the gyro scale factor error;

[0131] Μ g =[α xz α xy α yz α yx α zy α zx ] T is the gyro installation error;

[0132] v g is the gyro angular rate white noise;

[0133] δf s is the accelerometer error, and its error model is:

[0134]

[0135] is the constant bias of the accelerometer;

[0136] S a =[s ax s ay s az ] T is the accelerometer scale factor error;

[0137] Μ a =[β xz β xy β yz β yx β zy β zx ] T is the accelerometer installation error.

[0138] v ais the accelerometer comparative white noise.

[0139] F is the system matrix, according to the attitude and velocity error equations:

[0140]

[0141] diag() means converting a 3×1 vector into a 3×3 matrix with vector elements on the diagonal;

[0142] fun() means converting a 3×1 vector into the following 3×6 matrix;

[0143]

[0144] W is the white noise of the gyroscope and accelerometer.

[0145]

[0146] The observation equation is:

[0147] Z=HX+v

[0148] Among them, Z is the observation vector, and the velocity error is selected as the observation vector, that is, Z = δV n .

[0149] H is the observation matrix, then H=[0 3×3 Ι 3×3 0 3×24 ], v is the observation noise.

[0150] Exemplarily, the dual-axis rotating inertial navigation system rotates according to a preset eight-position rotation scheme, remains stationary for 5 minutes at each position, and the Kalman filter 1 performs state estimation during the stationary period. During the period when the dual-axis rotation mechanism rotates to the next position, the Kalman filter 1 skips the observation update and suspends the state estimation of the calibration parameters. When the dual-axis rotation mechanism completes the eight-position rotation scheme, the first stage ends, and the gyro constant drift, accelerometer constant bias, gyro scale factor error, accelerometer scale factor error, gyro installation error, and accelerometer installation error in the state vector are read as parameter calibration results.

[0151] For example, the inner lever arm calibrated in the second stage refers to the three vectors from the rotation center of the dual-axis indexing mechanism to the sensitive centers of the three accelerometers, corresponding to 9 parameters (see Figure 4 ). Three accelerometers A j The error vector between the sensitive point (j = X, Y, Z) and the rotation center O is represented by r X =[r Xx r Xy r Xz ] T、r Y =[r Yx r Yy r Yz ] T 、r Z =[r Zx r Zy r Zz ] T , which is the inner lever arm. X , U Y and U Z Represents the unit vector corresponding to each accelerometer sensitive axis.

[0152] When the dual-axis RINS is in angular motion relative to the inertial space, the accelerometer A j The output f of (j=X,Y,Z) j for:

[0153]

[0154] In the formula,

[0155] f O represents the specific force at point O, represents f O Along the sensitive axis U j projection;

[0156] ω is and represents the angular velocity and angular acceleration of the rotating coordinate system relative to the inertial coordinate system, and ω is =[ω x ω y ω z ] T ;

[0157]

[0158] (j = X, Y, Z) is the accelerometer specific force output error caused by the coupling of the inner lever arm and angular motion. Its vector form is:

[0159] δf=M vR r R

[0160] In the formula,

[0161]

[0162] r R =[(r X ) T (r Y ) T (r Z )T ] T

[0163] Exemplarily, in stage 2, the constructed Kalman filter 2 consists of a state equation and an observation equation.

[0164] The state equation is:

[0165]

[0166] in,

[0167] State Vector

[0168] They are attitude error and velocity error respectively;

[0169] Since the solution process of the dual-axis rotation RINS has been compensated using the parameter calibration results of the first stage during the second stage calibration, the attitude error and velocity error equations become:

[0170]

[0171] According to the attitude and velocity error equations, the system matrix F2 can be obtained:

[0172]

[0173] The observation equation is:

[0174] Z2=H2X2+v

[0175] Among them, Z2 is the observation vector, and the velocity error is selected as the observation vector, that is,

[0176] H2 is the observation matrix, then H2=[0 3×3 Ι 3×3 0 3×9 ].

[0177] For example, in the second stage, the definition of the six-order continuous indexing scheme performed by the dual-axis indexing mechanism is shown in Table 2. (See the schematic diagram Figure 5 )

[0178] Table 2 Definition of the six-order continuous transposition scheme

[0179]

[0180]

[0181] As an example, the observability of the inner lever arm under the six-order continuous indexing scheme is demonstrated below:

[0182] The observability analysis method of PWCS (Piece-Wise Constant System) uses the extracted observability matrix (Striping Observability Matrix, SOM) to analyze the observability of the system.

[0183] Since the PWCS observability analysis is independent of system noise and observation noise, the state equation and observation equation of the Kalman filter 2 can be discretized and simplified as:

[0184]

[0185] Where, F j 、H j are the state transition matrix and the observation matrix of the system within the j-th time interval respectively, and r is the number of selected time intervals.

[0186] Then, the observability matrix of the system is:

[0187]

[0188] The extracted observability matrix SOM is:

[0189]

[0190] When Rank(Q SOM (r)) = n, the system is completely observable.

[0191] When Rank(Q SOM (r)) < n, the system is incompletely observable.

[0192] Analyze the observability of the Kalman filter 2 under the six-order repositioning scheme. Select the sampling interval as 0.5 s. The rank of Q SOM (r) gradually increases with time and reaches full rank at the 7th time interval, that is, Rank(Q SOM (7)) = 15. It shows that the six-order repositioning scheme designed by the present invention can ensure that all state variables of the Kalman filter 2 are observable, and thus the calibration of 9 inner lever arm parameters can be realized by using the estimation of the Kalman filter 2.

[0193] Exemplarily, the two-axis rotary inertial navigation system rotates continuously according to the six-order repositioning scheme described above, and the constructed Kalman filter 2 is used for state estimation. Repeat the six-order repositioning scheme 20 times, for a total of 6 minutes. The second stage ends, and the estimated value of the inner lever arm in the state vector is read as the parameter calibration result.

Claims

1. A method for field self-calibration of a helicopter dual-axis rotating inertial navigation system, characterized in that: include: Step 1: Perform a power-on check on the dual-axis rotation inertial navigation system, and align it if the power-on check result is normal; Step 2, controlling the dual-axis indexing mechanism of the dual-axis rotation inertial navigation system to execute the eight-position indexing scheme, and performing navigation solution to obtain a first velocity error, using the first velocity error as the observation Z of the first Kalman filter, and using the first Kalman filter to estimate the gyro constant drift, the accelerometer constant bias, the gyro scale factor error, the accelerometer scale factor error, the gyro installation error, and the accelerometer installation error to obtain the first stage parameter calibration result; Step 3, the dual-axis indexing mechanism of the axis rotation inertial navigation system rotates periodically according to the six-order continuous indexing scheme, and performs navigation solution and parameter compensation using the first-stage parameter calibration results to obtain a second velocity error, and uses the second velocity error as the observation Z of the second Kalman filter, and uses the second Kalman filter to estimate the inner lever arm to obtain the second-stage parameter calibration results; Step 4: Bind the first-stage parameter calibration results and the second-stage parameter calibration results to the dual-axis rotation inertial navigation system.

2. The outfield self-calibration method for a helicopter dual-axis rotating inertial navigation system according to claim 1 is characterized in that: The eight-position indexing scheme includes: S11, placed in the initial position; in the initial position, x s The axis points east, y s The axis points north, z s The axis points to the sky; S12, around y s The axis rotates 270° counterclockwise; S13, around y s The axis rotates 180° clockwise; S14, around z s The axis rotates 270° counterclockwise; S15, around z s The axis rotates 180° clockwise; S16, around y s The axis rotates 180° clockwise; S17, go around z first s The axis rotates 45° clockwise; then rotates around y s The axis rotates 45° clockwise; S18, around y s The axis rotates 180° clockwise; x s Axis, y s Axis, z s The axes are the ideal sensitive axis directions of the x-axis gyro, y-axis gyro and z-axis gyro, and the outer rotation axis and inner rotation axis of the double-axis indexing mechanism are respectively s Axis and z s Axis coincidence; Ox s y s z s The origin O of the coordinate system is the rotation center of the double-axis indexing mechanism.

3. The outfield self-calibration method for a helicopter dual-axis rotating inertial navigation system according to claim 1 is characterized in that: The six-order continuous transposition scheme includes: S21, around y s Axis and z s The axes simultaneously perform uniform angular motion at a first preset angular acceleration rate and continue for a preset duration; S22, around y s The axis makes a uniform angular motion at the first angular velocity; at the same time, around the z s The axis performs a uniform deceleration angular motion at a first preset angular acceleration rate and continues for a preset time period; S23, around y s The axis performs a uniform deceleration angular motion at a first preset angular acceleration rate and continues for a preset time period; S24, around z s The axis performs uniform angular motion at a first preset angular acceleration rate and continues for a preset time period; S25, around y s The axis performs uniform acceleration angular motion at a first preset angular acceleration rate; at the same time, around the z s The axis performs uniform angular motion at a first angular rate and continues for a preset time. S26, around y s Axis and z s The axes simultaneously perform uniform deceleration angular motion at a first preset angular acceleration rate and continue for a preset duration.

4. The outfield self-calibration method for a helicopter dual-axis rotating inertial navigation system according to claim 3 is characterized in that: The first preset angular acceleration The first angular velocity is determined according to a first preset angular acceleration rate and a preset duration.

5. The outfield self-calibration method for a helicopter dual-axis rotating inertial navigation system according to claim 1 is characterized in that: The first Kalman filter includes a first state equation and a first observation equation; The first state equation is Where X is the first state vector, φ n ,δV n They are the first attitude error and the first velocity error respectively; The first attitude error and first velocity error equations of the inertial navigation system are: In the formula, n, s, i, e in superscript and / or subscript represent navigation coordinate system, rotation coordinate system, geocentric inertial coordinate system, and earth coordinate system, respectively; Represents the direction cosine matrix from the rotation coordinate system to the navigation coordinate system; is the Earth's rotation angular velocity, is the angular velocity caused by the motion of the carrier; f represents the specific force output of the accelerometer; δω s is the gyro error, and its error model is: is the gyro constant drift; S g =[s gx s gy s gz ] T is the gyro scale factor error; is the angular velocity of the rotating coordinate system relative to the inertial coordinate system; Μ g = [α xz α xy α yz α yx α zy α zx T is the gyro installation error;​ v g is the gyro angular rate white noise; δf s is the accelerometer error, and its error model is: is the constant bias of the accelerometer; S a =[s ax s ay s az ] T is the accelerometer scale factor error; Μ a =[β xz β xy β yz β yx β zy β zx ] T is the accelerometer installation error. v a is the accelerometer comparative white noise; F is the first system matrix, according to the attitude and velocity error equations: diag() means converting a 3×1 vector into a 3×3 matrix with vector elements on the diagonal; fun() means converting a 3×1 vector into the following 3×6 matrix; W is the white noise of the gyroscope and accelerometer; The first observation equation is: Z = HX + v; Where Z is the first observation vector, Z = δV n ; H is the first observation matrix, H = [0 3×3 Ι 3×3 0 3×24 ], v is the observation noise.

6. The method for outfield self-calibration of a helicopter dual-axis rotating inertial navigation system according to claim 5, characterized in that: The second Kalman filter includes a second state equation and a second observation equation; The second state equation is: in, The second state vector They are the second attitude error and the second velocity error respectively; According to the attitude and velocity error equations, the system matrix F2 can be obtained: The second observation equation is: Z2=H2X2+v Among them, Z2 is the observation vector, H2 is the observation matrix, H2=[0 3×3 Ι 3×3 0 3×9 ]; r R ＝[(r X ) T (r Y ) T (r Z ) T ] T ； r R is the inner lever arm, r X 、r Y 、r Z They represent the error vectors in three directions between the sensitive points of the three accelerometers and the rotation center point of the dual-axis indexing mechanism, r X =[r Xx r Xy r Xz ] T 、r Y =[r Yx r Yy r Yz ] T 、rZ=[r Zx r Zy r Zz ] T ; δf=M vR r R ; ω is and Represents the angular velocity and angular acceleration of the rotating coordinate system relative to the inertial coordinate system, ω is =[ω x ω y ω z ] T ; U X , U Y and U Z Represents the unit vector corresponding to each accelerometer's sensitive axis.

7. The method for outfield self-calibration of a helicopter dual-axis rotating inertial navigation system according to claim 1, characterized in that: Before step 4, the method further comprises: Determine whether the difference between the first-stage parameter calibration result and the second-stage parameter calibration result and the historical parameter calibration result is within a preset range.

8. The method for outfield self-calibration of a helicopter dual-axis rotating inertial navigation system according to claim 1, characterized in that: Before step 3, the method further comprises: Determine that the difference between the first-stage parameter calibration result and the corresponding historical parameter calibration result is within a preset range; Correspondingly, before step 4, the method further includes: Determine whether the difference between the second-stage parameter calibration result and the corresponding historical parameter calibration result is within a preset range.

Images