WO2020062807A1 - 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法 - Google Patents

改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法 Download PDF

Info

Publication number
WO2020062807A1
WO2020062807A1 PCT/CN2019/079136 CN2019079136W WO2020062807A1 WO 2020062807 A1 WO2020062807 A1 WO 2020062807A1 CN 2019079136 W CN2019079136 W CN 2019079136W WO 2020062807 A1 WO2020062807 A1 WO 2020062807A1
Authority
WO
WIPO (PCT)
Prior art keywords
error
equation
measurement
sins
state
Prior art date
Application number
PCT/CN2019/079136
Other languages
English (en)
French (fr)
Inventor
徐晓苏
杨阳
袁杰
翁铖铖
梁紫依
刘兴华
Original Assignee
东南大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 东南大学 filed Critical 东南大学
Publication of WO2020062807A1 publication Critical patent/WO2020062807A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20Instruments for performing navigational calculations
    • G01C21/203Specially adapted for sailing ships
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/005Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/10Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
    • G01C21/16Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
    • G01C21/165Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation combined with non-inertial navigation instruments
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/10Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
    • G01C21/16Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
    • G01C21/18Stabilised platforms, e.g. by gyroscope

Definitions

  • the invention belongs to the field of inertial navigation and relates to an application method of an improved unscented Kalman filter algorithm in underwater integrated navigation SINS / DVL / GPS.
  • Autonomous Underwater Vehicles have a wide range of applications in civilian and military applications.
  • Military aspects such as theater reconnaissance, detection and removal of mines, submarine countermeasures, maritime early warning blockades or ports, attacks on enemy ships or submarines, destruction of oil facilities and communication networks, underwater relay communications, etc .; civilian aspects such as marine resource survey and development, Marine rescue and salvage.
  • navigation remains one of the main technical challenges facing AUVs.
  • the navigation system must provide accurate attitude, heading, speed, and position information during long voyages, and accurate navigation capabilities are a key technology for the effective application and safe recovery of AUVs.
  • due to its size, weight, power supply restrictions, and the particularity and concealment of the water medium it is a difficult task to achieve accurate navigation of the AUV.
  • Inertial navigation is an autonomous navigation that does not require contact with the outside world and has good concealment. It is an ideal navigation method for AUVs.
  • the inertial navigation system has the problem of error accumulation.
  • the inertial navigation system can maintain good navigation accuracy in a short time, but the error accumulation effect is very obvious with the extension of time, and the navigation accuracy cannot be guaranteed. Therefore, other auxiliary navigation systems are usually used to correct the cumulative error of inertial navigation.
  • Doppler velocimetry Roland C and other methods. Due to its high navigation accuracy and no error accumulation, GPS is the preferred method for high-precision navigation and positioning.
  • the present invention discloses an application method of an improved unscented Kalman filtering algorithm in underwater integrated navigation, which can improve the filtering settlement efficiency and navigation accuracy of an underwater integrated navigation system, and ensure real-time performance and stability.
  • the present invention provides the following technical solutions:
  • the application method of the improved unscented Kalman filter algorithm in underwater integrated navigation includes the following steps:
  • the coordinate system established in the step (1) specifically includes:
  • inertial coordinate system does not rotate with the earth, the origin is located at the center of the earth, the z i axis points to the north pole, the x i axis points to the vernal equinox, and the y i axis and x i and z i form a right-handed coordinate system;
  • Earth coordinate system fixedly connected to the earth, the origin is located at the center of the earth, the x e axis crosses the intersection of the prime meridian and the equator, the z e axis points to the north pole, and the y e axes x e , z e form a right-handed coordinate system;
  • carrier coordinate system the origin is located at the center of the carrier, the z b axis is perpendicular to the carrier upwards, x b points to the front of the carrier, and y b and x b and z b form a right-handed coordinate system;
  • n the navigation coordinate system that coincides with the east-north-sky geographic coordinate system.
  • step (2) establishing a state equation of the SINS / DVL subsystem in the underwater submarine phase specifically includes the following steps:
  • attitude error angle ( ⁇ E ⁇ N ⁇ U )
  • gyro constant value drift ( ⁇ bx ⁇ by ⁇ bz )
  • random constant error of the accelerometer As the state quantity of the SINS system, it is recorded as:
  • the speed error equation represented by quaternion is derived from the speed differential equation of the ideal inertial specific gravity equation and the actual speed differential equation of the strapdown inertial navigation system:
  • V n is the ideal speed of the carrier in the n series
  • f b is the specific force in the b system
  • Is the quaternion of the conversion from p to n Is the quaternion of the conversion from b to p, with Represent the transformation matrices of p-series to n-series and b-series to p-series
  • represents the actual measured value of the carrier
  • represents the error between the ideal value of the carrier and the actual measured value
  • g n is the gravity acceleration under the n system. Is the accelerometer error vector under b system;
  • R M and R N respectively represent the radius of curvature of the meridian circle and the radius of curvature of the circle of circle, L is the latitude of the carrier, ⁇ is the longitude of the carrier, and h is the height of the carrier;
  • F N [ ⁇ ] is a non-linear continuous function
  • ⁇ d represents the reciprocal of the correlation time of the speed offset error
  • ⁇ d represents the excitation white noise
  • the state vector of the integrated navigation system is The noise vector of the system is The equation of state of the system is expressed as:
  • the state function F 1 [ ⁇ ] is a nonlinear continuous function
  • ⁇ 1 (t) is the noise matrix of the subsystem.
  • step (3) of establishing the measurement equation of the SINS / DVL subsystem in the underwater submarine phase specifically includes the following steps:
  • the posture matrix used for the change is provided by SINS v EI , v NI , v UI respectively represent the three-axis velocity calculated by SINS, v ED , v ND , v UD respectively represent the three-axis velocity measured by DVL, v E , v N , v U represent the carrier in the navigation coordinate system
  • the real speed under vx , v y , v z represents the real speed in the carrier coordinate system, ⁇ v DE , ⁇ v DN , ⁇ v DU is the error after the DVL speed measurement error is converted to the navigation coordinate system;
  • H 1 [ ⁇ ] is a non-linear continuous function
  • V 1 (t) is the measurement noise
  • step (4) establishing a state equation of the SINS / GPS subsystem at the surface position correction stage specifically includes the following steps:
  • the noise error of the system is:
  • F 2 [ ⁇ ] is a nonlinear continuous function
  • ⁇ 2 (t) is the noise matrix of the subsystem.
  • step (5) of establishing the measurement equation of the SINS / GPS subsystem in the surface position correction stage specifically includes the following steps:
  • L, ⁇ , V E and V N are the positions and velocities calculated by the SINS solution
  • L G , ⁇ G , V GE and V GN are the positions, velocities of the GPS output
  • ⁇ ⁇ represents the corresponding error
  • H 2 [ ⁇ ] is a non-linear continuous function
  • V 2 is a measurement noise
  • the non-linear filtering equation of the underwater submarine stage established by integrating the steps (2) and (3) is:
  • step (6) specifically includes the following steps:
  • X k and Z k respectively, for the system state vector and the measurement vector, W k and V k t k are the time of noise and measurement noise matrix array subsystem stage submerged underwater, are zero mean and,
  • the statistical characteristics are as follows:
  • Q k and R k are the subsystem noise covariance matrix and the measurement noise covariance matrix respectively; the specific algorithm process is as follows:
  • ⁇ i, k / k-1 f 1 ( ⁇ i, k-1 )
  • ⁇ i, k / k-1 is the i-th sample point predicted at t k ;
  • Z i, k / k-1 is the i-th measurement value
  • x k and z k are respectively the state vector and the measurement vector of the system at time t k
  • w k and v k are the noise matrix and the measurement noise matrix of the subsystem at the surface position correction stage, respectively, and the mean is zero.
  • the statistical characteristics are as follows: q k and r k are the noise covariance matrix and the measured noise covariance matrix of the subsystem;
  • ⁇ i, k / k-1 f 2 ( ⁇ i, k-1 )
  • ⁇ i, k / k-1 is the i-th sample point predicted at t k ;
  • Z i, k / k-1 is the i-th measurement value
  • the position information of the filtering result every time the AUV rises to the surface is used as the new position information for the next submarine voyage, and the position is periodically corrected.
  • the present invention has the following advantages and beneficial effects:
  • the application method of the improved unscented Kalman filtering method proposed in the present invention in integrated navigation simplifies the UKF navigation algorithm based on the spherically distributed simplex sampling of complex additive noise, greatly reducing the dimension of the system state vector and the filtering algorithm.
  • the complexity of the calculation has good real-time, stability and accuracy.
  • FIG. 1 is a schematic diagram of the integrated system navigation for implementing the method of the present invention.
  • Fig. 2 is a working flowchart of the integrated navigation system.
  • Figure 3 is a schematic diagram of the UT transformation used in the improved unscented Kalman filtering method.
  • Figure 4 is a block diagram of the solution flow of the improved unscented Kalman filtering method.
  • Step 1 Define the coordinate system to be used
  • inertial coordinate system does not rotate with the earth, the origin is located at the center of the earth, the z i axis points to the north pole, the x i axis points to the vernal equinox, and the y i axis and x i and z i form a right-handed coordinate system;
  • Earth coordinate system fixedly connected to the earth, the origin is located at the center of the earth, the x e axis crosses the intersection of the prime meridian and the equator, the z e axis points to the north pole, and the y e axes x e , z e form a right-handed coordinate system;
  • carrier coordinate system the origin is located at the center of the carrier, the z b axis is perpendicular to the carrier upwards, x b points to the front of the carrier, and y b and x b and z b form a right-handed coordinate system;
  • n the navigation coordinate system that coincides with the east-north-sky geographic coordinate system.
  • Step 2 Select state variables and measurements to establish a non-linear system model in the underwater submarine phase; take attitude error angle ( ⁇ E ⁇ N ⁇ U ), speed error ( ⁇ v E ⁇ v N ⁇ v U ), position error ( ⁇ L ⁇ ⁇ h) , Gyro constant drift ( ⁇ bx ⁇ by ⁇ bz ) and random constant error of accelerometer
  • attitude error angle ⁇ E ⁇ N ⁇ U
  • speed error ⁇ v E ⁇ v N ⁇ v U
  • position error ⁇ L ⁇ ⁇ h
  • Gyro constant drift ⁇ bx ⁇ by ⁇ bz
  • random constant error of accelerometer As the state quantity of the SINS system, it is recorded as:
  • the speed error equation represented by quaternion is derived from the speed differential equation of the ideal inertial specific gravity equation and the actual speed differential equation of the strapdown inertial navigation system:
  • V n is the ideal speed of the carrier in the n series
  • f b is the specific force in the b system
  • Is the quaternion of the conversion from p to n Is the quaternion of the conversion from b to p, with Represent the transformation matrices of p-series to n-series and b-series to p-series, respectively.
  • represents the actual measured value of the carrier
  • represents the error between the ideal value of the carrier and the actual measured value
  • g n is the gravity acceleration under the n system. Is the accelerometer error vector for the b system.
  • R M and R N respectively indicate the radius of curvature of the meridian circle and the radius of curvature of the cymbal circle, and h represents the height at which the carrier is located.
  • F N [ ⁇ ] is a non-linear continuous function.
  • ⁇ d represents the reciprocal of the correlation time of the speed offset error
  • ⁇ d represents the excitation white noise
  • the state vector of the integrated navigation system is The noise vector of the system is The equation of state of the system can be expressed as:
  • the state function F 1 [ ⁇ ] is a nonlinear continuous function
  • ⁇ 1 (t) is the noise matrix of the subsystem.
  • DVL measures the component of ground speed in the coordinate system of the carrier, in order to measure the speed of its output and the speed formed by SINS, the output speed of DVL must be transformed into the navigation coordinate system. among them,
  • H 1 [ ⁇ ] is a non-linear continuous function
  • V 1 (t) is the measurement noise
  • Step 3 Select the state quantity and quantity measurement to establish a non-linear system model at the surface position correction stage, and take the state variables:
  • the noise error of the system is:
  • F 2 [ ⁇ ] is a nonlinear continuous function
  • ⁇ 2 (t) is the noise matrix of the subsystem.
  • H 2 [ ⁇ ] is a non-linear continuous function
  • V 2 is a measurement noise
  • Step 4 Discretize the equations established above and perform the unscented Kalman filter solution improved according to the algorithm shown in FIG. 4 to implement time update, measurement update, and filter update.
  • the improved Kalman filter algorithm is introduced as follows:
  • SSUT Spherical distribution simplex sampling transformation
  • Unscented Kalman (UKF) inference algorithm it is generally necessary to perform state augmentation on system noise and measurement noise, but when the system noise and measurement noise are additive noise, it is not necessary to do augmentation processing, which is beneficial to further Reduce filtering calculations.
  • the present invention studies a simplified SSUT sampling UKF algorithm based on complex additive noise.
  • the nonlinear discrete-time system model of complex additive noise can be expressed as:
  • the specific algorithm flow is as follows:
  • ⁇ i, k / k-1 f ( ⁇ i, k-1 )
  • the Kalman filter equation established in the underwater submarine phase is discretized.
  • X k and Z k respectively, for the system state vector and the measurement vector, W k and V k t k are the time of noise and measurement noise matrix array subsystem stage submerged underwater, are zero mean and,
  • the statistical characteristics are as follows:
  • Q k and R k are the subsystem noise covariance matrix and the measurement noise covariance matrix, respectively.
  • ⁇ i, k / k-1 f 1 ( ⁇ i, k-1 )
  • x k and z k are respectively the state vector and the measurement vector of the system at time t k
  • w k and v k are the noise matrix and the measurement noise matrix of the subsystem at the surface position correction stage, respectively, and the mean is zero.
  • the statistical characteristics are as follows: q k and r k are the noise covariance matrix and the measured noise covariance matrix of the subsystem, respectively.
  • the specific algorithm flow is as follows:
  • ⁇ i, k / k-1 f 2 ( ⁇ i, k-1 )
  • the position information of the filtering result when the AUV rises to the surface each time is used as the new position information of the next submarine, so as to realize the timing correction of the position and overcome the cumulative error of the inertial navigation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Automation & Control Theory (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Navigation (AREA)
  • Position Fixing By Use Of Radio Waves (AREA)

Abstract

一种改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,包括:对需要用到的坐标系进行定义;建立水下潜航阶段SINS/DVL子***状态方程;建立水下潜航阶段SINS/DVL子***量测方程;建立水面位置修正阶段SINS/GPS子***的状态方程;建立水面位置修正阶段SINS/GPS子***的量测方程;建立水下潜航阶段的非线性滤波方程,在水下潜航一段时间以后,AUV浮出水面,建立水面位置修正阶段非线性滤波方程,进行改进的无迹卡尔曼滤波解算,完成时间更新、量测更新和滤波更新,完成定时位置信息校正。该方法能够提高水下组合导航***的滤波解算效率及导航精度,降低了滤波算法计算的复杂程度,保证实时性和稳定性。

Description

改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法 技术领域
本发明属于惯性导航领域,涉及一种改进的无迹卡尔曼滤波算法在水下组合导航SINS/DVL/GPS中的应用方法。
背景技术
水下自主航行器(Autonomous Underwater Vehicle,AUV)无论在民用还是军事方面都有广泛的应用。军事方面如战区侦察、探测扫除水雷、潜艇对抗、海上预警封锁航线或港口、攻击敌舰船或潜艇、破坏石油设施及通讯网络、水下中继通讯等;民用方面如海洋资源勘察与开发、海洋救险和打捞等。就目前的发展而言,导航问题仍然是AUV所面临的主要技术挑战之一。导航***必须提供长航时的精确姿态、航向、速度和位置信息,而且精确的导航能力是AUV有效应用和安全回收的一个关键技术。但由于受其大小、重量、电源使用的限制及水介质的特殊性、隐蔽性等因素的影响,实现AUV的精确导航是一项艰难的任务。
远航程AUV因为其远航程的特点,要求长时间、长距离的在水下航行,这就要求导航必需满足一定的精度。惯性导航是一种自主导航,不需要与外界的联系,隐蔽性好,是AUV理想的导航方式。但是,惯导有误差积累的问题,短时间内惯导可以保持很好的导航精度,但是随着时间的延长,误差积累效应非常明显,无法保证导航精度。因此,通常都是采用其它的辅助导航***配合以修正惯导的累积误差。如多普勒测速,罗兰C等方法。GPS由于具有导航精度高,没有误差积累的特点,是高精度导航定位的首选方法。但是对于水下航行器而言,需要浮出水面才能接收GPS信号,而且为了隐蔽不能长时间漂浮在水面,因此GPS只能作为校正***使用。此外,由于AUV组合导航***在本质上具有非线性特性,传统上采用扩展卡尔曼滤波(EKF)对非线性模型进行线性化处理,在此过程中由于忽略了模型的高阶导数,随着时间的增长,导航精度难以得到保证,而且在计算过程中反复计算***方程的雅各比矩阵,增加了滤波计算量。
发明内容
为解决上述问题,本发明公开了改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,能够提高水下组合导航***的滤波结算效率及导航精度,保证实时性和稳定性。
为了达到上述目的,本发明提供如下技术方案:
改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,包括如下步骤:
(1)对需要用到的坐标系进行定义:
(2)建立水下潜航阶段SINS/DVL子***的状态方程:以SINS的3轴姿态误差,3轴速度误差,位置误差,3轴陀螺仪随机漂移、3轴加速度计零偏和DVL速度偏移误差、刻度系数误差建立21维状态向量,根据***误差方程建立该子***的状态方程;
(3)建立水下潜航阶段SINS/DVL子***的量测方程:根据SINS解算的3维速度量和DVL测量的速度量之差作为量测量,并结合步骤(2)选取的误差状态向量建立该子***的量测方程;
(4)建立水面位置修正阶段SINS/GPS子***的状态方程:不考虑天向运动速度及误差,以SINS的3轴姿态误差、速度误差、位置误差、3轴陀螺仪随机漂移、加速度计零偏建立12维状态向量,根据误差方程建立该子***的状态方程;
(5)建立水面位置修正阶段SINS/GPS子***的量测方程:以SINS解算出来的位置、速度与GPS输出的位置、速度之差作为量测量,并结合步骤(4)选取的误差状态向量建立该子***的量测方程;
(6)综合步骤(2)和步骤(3)建立水下潜航阶段的非线性滤波方程,在水下潜航一段时间以后,AUV浮出水面,结合步骤(4)和步骤(5)建立水面位置修正阶段非线性滤波方程,进行改进的无迹卡尔曼滤波解算,完成时间更新、量测更新和滤波更新,完成定时位置信息校正。
进一步的,所述步骤(1)建立的坐标系具体包括:
i——惯性坐标系:不随地球旋转,原点位于地球中心,z i轴指向北极,x i轴指向春分点,y i轴与x i、 z i构成右手坐标系;
e——地球坐标系:与地球固联,原点位于地心,x e轴穿越本初子午线与赤道交点,z e轴指向北极,y e轴x e、z e构成右手坐标系;
b——载体坐标系:原点位于运载体中心,z b轴垂直运载体向上,x b指向运载体前方,y b与x b、z b构成右手坐标系;
p——实际计算得出的平台坐标系;
n——与东-北-天地理坐标系重合的导航坐标系。
进一步的,所述步骤(2)中建立水下潜航阶段SINS/DVL子***的状态方程具体包括如下步骤:
取姿态误差角(φ E φ N φ U)、速度误差(δv E δv N δv U)位置误差(δL δλ δh)、陀螺常值漂移(ε bx ε by ε bz)和加速度计随机常值误差
Figure PCTCN2019079136-appb-000001
作为SINS***的状态量,记为:
Figure PCTCN2019079136-appb-000002
通过理想惯导比力方程的速度微分方程和捷联惯导***的实际速度微分方程,推导出四元数表示的速度误差方程:
Figure PCTCN2019079136-appb-000003
其中,V n是载体在n系下的理想速度,
Figure PCTCN2019079136-appb-000004
是n系下的地球自转角速度,
Figure PCTCN2019079136-appb-000005
为n系相对于e系的角速度在n系下的投影,f b为b系下的比力;式中,
Figure PCTCN2019079136-appb-000006
为p系到n系的转换四元数,
Figure PCTCN2019079136-appb-000007
为b系到p系的转换四元数,
Figure PCTCN2019079136-appb-000008
Figure PCTCN2019079136-appb-000009
分别代表p系到n系以及b系到p系的转换矩阵;
Figure PCTCN2019079136-appb-000010
式中,“~”表示载体实际测量值,δ表示载体的理想值与实际测量值之间的误差,g n为n系下的重力加速度,
Figure PCTCN2019079136-appb-000011
为b系下的加速度计误差向量;
四元数姿态误差方程:
Figure PCTCN2019079136-appb-000012
其中,
Figure PCTCN2019079136-appb-000013
表示n系相对于i系的旋转角速度在n系下的投影,
Figure PCTCN2019079136-appb-000014
为n系到p系的方向余弦矩阵,ε b为陀螺仪误差向量在b系下的投影,B为关于四元数的4×3维矩阵;
位置误差方程:
Figure PCTCN2019079136-appb-000015
Figure PCTCN2019079136-appb-000016
Figure PCTCN2019079136-appb-000017
其中,R M和R N分别表示地球的子午圈曲率半径和卯酉圈曲率半径,L表示载体的纬度,λ表示载体的经度,h表示载体的高度;
SINS***的噪声:
W N(t)=[ω gx ω gy ω gz ω ax ω ay ω az] T
则SINS的***误差方程可以表示为:
Figure PCTCN2019079136-appb-000018
F N[·]为非线连续函数;
取DVL速度偏移误差(δV dx δV dy δV dz)和刻度系数误差(Δk dx Δk dy Δk dz)作为DVL***状态变量,记为:
X D(t)=[δV dx δV dy δV dz δk dx δk dy δk dz] T
DVL的误差模型为:
Figure PCTCN2019079136-appb-000019
Figure PCTCN2019079136-appb-000020
其中,β d表示速度偏移误差的相关时间倒数,ω d表示激励白噪声;
相应的误差状态方程为:
Figure PCTCN2019079136-appb-000021
式中:
Figure PCTCN2019079136-appb-000022
W D(t)=[ω dx ω dy ω dz] T;ω di(i=x,y,z)为激励白噪声;G D(t)=[I 3×3 O 3×3];τ di(i=x,y,z)表示速度偏移误差的相关时间;
选取SINS和DVL子***的误差状态变量,则组合导航***的状态向量为
Figure PCTCN2019079136-appb-000023
***的噪声向量为
Figure PCTCN2019079136-appb-000024
***的状态方程表示为:
Figure PCTCN2019079136-appb-000025
式中,状态函数F 1[·]为非线性连续函数,Γ 1(t)为该子***噪声阵。
进一步的,所述步骤(3)中建立水下潜航阶段SINS/DVL子***的量测方程具体包括如下步骤:
由SINS和DVL形成的量测量:
Figure PCTCN2019079136-appb-000026
其中,由SINS提供变化所用的姿态矩阵
Figure PCTCN2019079136-appb-000027
v EI,v NI,v UI分别表示SINS解算出的三轴速度,v ED,v ND,v UD分别表示DVL测出的三轴速度,v E,v N,v U表示载体在导航坐标系下的真实速度,v x,v y,v z表示在载体坐标系下的真实速度,δv DE,δv DN,δv DU为DVL测速误差换算到导航坐标系后的误差;
将上式展开,并结合之前选取的***误差状态
Figure PCTCN2019079136-appb-000028
得该组合***的量测方程为:
Z 1=H 1[X 1(t),t]+V 1(t)
式中:H 1[·]为非线性连续函数;V 1(t)为量测噪声。
进一步的,所述步骤(4)中建立水面位置修正阶段SINS/GPS子***的状态方程具体包括如下步骤:
取状态变量:
Figure PCTCN2019079136-appb-000029
***的噪声误差为:
W 2(t)=[0 0 ω gx ω gy ω gz ω ax ω ay 0 0 0 0 0] T
建立基于SINS/GPS的AUV组合导航连续***状态方程:
Figure PCTCN2019079136-appb-000030
式中:F 2[·]为非线性连续函数,Γ 2(t)为该子***噪声阵。
进一步的,所述步骤(5)中建立水面位置修正阶段SINS/GPS子***的量测方程具体包括如下步骤:
将SINS解算出来的位置、速度与GPS输出的位置位置、速度之差作为AUV水面位置修正阶段滤波解算的量测方程:
Figure PCTCN2019079136-appb-000031
式中L、λ、V E和V N分别为SINS解算出来的位置、速度,L G、λ G、V GE和V GN分别为GPS输出的位置位置、速度,δ·表示相应的误差;将上式展开,并结合之前选取的***误差状态X 2(t),得到量测方程为:
Z 2=H 2[X 2(t),t]+V 2(t)
式中:H 2[·]为非线性连续函数,V 2为量测噪声。
进一步的,所述步骤(6)中综合步骤(2)和步骤(3)建立的水下潜航阶段的非线性滤波方程为:
Figure PCTCN2019079136-appb-000032
结合步骤(4)和步骤(5)建立的水面位置修正阶段非线性滤波方程为:
Figure PCTCN2019079136-appb-000033
进一步的,所述步骤(6)中进行改进的无迹卡尔曼滤波解算的过程具体包括如下步骤:
对水下潜航阶段的非线性滤波方程离散化有:
Figure PCTCN2019079136-appb-000034
其中,X k和Z k分别为***在t k时刻的状态向量以及量测向量,W k和V k分别为水下潜航阶段子***的噪 声阵和量测噪声阵,且均值均为零,统计特性如下:
Figure PCTCN2019079136-appb-000035
Q k和R k分别为子***噪声协方差阵和量测噪声协方差阵;具体的算法过程如下:
<1>初始化增广状态向量及估计误差方差
Figure PCTCN2019079136-appb-000036
Figure PCTCN2019079136-appb-000037
<2>计算sigma点χ i,k-1和相应的加权因子W i
Figure PCTCN2019079136-appb-000038
W i=(1-W 0)/22,其中,0≤W 0≤1
其中,
Figure PCTCN2019079136-appb-000039
表示21维状态量的第i个采样点;
<3>时间更新方程,得到一步预测
Figure PCTCN2019079136-appb-000040
和一步预测误差协方差P k/k-1
χ i,k/k-1=f 1i,k-1)
Figure PCTCN2019079136-appb-000041
Figure PCTCN2019079136-appb-000042
χ i,k/k-1为t k时刻预测的第i个样本点;
<4>量测更新方程,得到k时刻的量测预测
Figure PCTCN2019079136-appb-000043
量测预测协方差
Figure PCTCN2019079136-appb-000044
和状态量与量测量之间的协方差
Figure PCTCN2019079136-appb-000045
Z i,k/k-1=h 1i,k/k-1)
Figure PCTCN2019079136-appb-000046
Figure PCTCN2019079136-appb-000047
Figure PCTCN2019079136-appb-000048
其中,Z i,k/k-1为第i个量测值;
<5>滤波更新方程,得到滤波增益矩阵K k、状态量的最优滤波估计
Figure PCTCN2019079136-appb-000049
和最优滤波估计误差协方差矩阵P k
Figure PCTCN2019079136-appb-000050
Figure PCTCN2019079136-appb-000051
Figure PCTCN2019079136-appb-000052
对水面位置修正阶段非线性滤波方程进行离散化有:
Figure PCTCN2019079136-appb-000053
其中,x k和z k分别为***在t k时刻的状态向量以及量测向量,w k和v k分别为水面位置修正阶段子***的噪声阵和量测噪声阵,且均值均为零,统计特性如下:
Figure PCTCN2019079136-appb-000054
Figure PCTCN2019079136-appb-000055
q k和r k分别为该子***噪声协方差阵和量测噪声协方差阵;
具体算法步骤如下:
<1>初始化增广状态向量及估计误差方差
Figure PCTCN2019079136-appb-000056
Figure PCTCN2019079136-appb-000057
<2>计算sigma点χ i,k-1和相应的加权因子W i
Figure PCTCN2019079136-appb-000058
W i=(1-W 0)/13,其中,0≤W 0≤1
其中,
Figure PCTCN2019079136-appb-000059
表示12维状态量的第i个采样点;
<3>时间更新方程,得到一步预测
Figure PCTCN2019079136-appb-000060
和一步预测误差协方差P k/k-1
χ i,k/k-1=f 2i,k-1)
Figure PCTCN2019079136-appb-000061
Figure PCTCN2019079136-appb-000062
χ i,k/k-1为t k时刻预测的第i个样本点;
<4>量测更新方程,得到k时刻的量测预测
Figure PCTCN2019079136-appb-000063
量测预测协方差
Figure PCTCN2019079136-appb-000064
和状态量与量测量之间的协方差
Figure PCTCN2019079136-appb-000065
Z i,k/k-1=h 2i,k/k-1)
Figure PCTCN2019079136-appb-000066
Figure PCTCN2019079136-appb-000067
Figure PCTCN2019079136-appb-000068
其中,Z i,k/k-1为第i个量测值;
<5>滤波更新方程,得到滤波增益矩阵K k、状态量的最优滤波估计
Figure PCTCN2019079136-appb-000069
和最优滤波估计误差协方差矩阵P k
Figure PCTCN2019079136-appb-000070
Figure PCTCN2019079136-appb-000071
Figure PCTCN2019079136-appb-000072
将每次AUV浮出水面的时的滤波结果的位置信息作为下一次潜航的新的位置信息,定时修正位置。
与现有技术相比,本发明具有如下优点和有益效果:
本发明提出的改进的无迹卡尔曼滤波方法在组合导航中的应用方法,基于复杂加性噪声的球面分布单形采样简化UKF导航算法,大大减少了***状态向量的维数,降低了滤波算法计算的复杂程度,具备较好的实时性、稳定性和精确度。
附图说明
图1为实现本发明方法的组合***导航原理图。
图2是组合导航***的工作流程图。
图3是改进的无迹卡尔曼滤波方法所用到的UT变换原理图。
图4是改进的无迹卡尔曼滤波方法的解算流程框图。
具体实施方式
以下将结合具体实施例对本发明提供的技术方案进行详细说明,应理解下述具体实施方式仅用于说明本发明而不用于限制本发明的范围。
本发明提供的一种改进的无迹卡尔曼滤波算法在水下组合导航***SINS/DVL/GPS中的应用方法,实现原理如图1-图4所示,其流程主要包括以下步骤:
步骤一:定义需要用到的坐标系;
(1)对需要用到的坐标系做如下定义:
i——惯性坐标系:不随地球旋转,原点位于地球中心,z i轴指向北极,x i轴指向春分点,y i轴与x i、z i构成右手坐标系;
e——地球坐标系:与地球固联,原点位于地心,x e轴穿越本初子午线与赤道交点,z e轴指向北极,y e轴x e、z e构成右手坐标系;
b——载体坐标系:原点位于运载体中心,z b轴垂直运载体向上,x b指向运载体前方,y b与x b、z b构成右手坐标系;
p——实际计算得出的平台坐标系;
n——与东-北-天地理坐标系重合的导航坐标系。
步骤二:选取状态变量和量测量建立水下潜航阶段的非线性***模型;取姿态误差角(φ E φ N φ U)、速度误差(δv E δv N δv U)位置误差(δL δλ δh)、陀螺常值漂移(ε bx ε by ε bz)和加速度计随机常值误差
Figure PCTCN2019079136-appb-000073
作为SINS***的状态量,记为:
Figure PCTCN2019079136-appb-000074
通过理想惯导比力方程的速度微分方程和捷联惯导***的实际速度微分方程,推导出四元数表示的速度误差方程:
Figure PCTCN2019079136-appb-000075
其中,V n是载体在n系下的理想速度,
Figure PCTCN2019079136-appb-000076
是n系下的地球自转角速度,
Figure PCTCN2019079136-appb-000077
为n系相对于e系的角速度在n系下的投影,f b为b系下的比力;式中,
Figure PCTCN2019079136-appb-000078
为p系到n系的转换四元数,
Figure PCTCN2019079136-appb-000079
为b系到p系的转换四元数,
Figure PCTCN2019079136-appb-000080
Figure PCTCN2019079136-appb-000081
分别代表p系到n系以及b系到p系的转换矩阵。
Figure PCTCN2019079136-appb-000082
式中,“~”表示载体实际测量值,δ表示载体的理想值与实际测量值之间的误差,g n为n系下的重力加速度,
Figure PCTCN2019079136-appb-000083
为b系下的加速度计误差向量。
四元数姿态误差方程:
Figure PCTCN2019079136-appb-000084
其中,
Figure PCTCN2019079136-appb-000085
表示n系相对于i系的旋转角速度在n系下的投影,
Figure PCTCN2019079136-appb-000086
为n系到p系的方向余弦矩阵,ε b为陀螺仪误差向量在b系下的投影,B为关于四元数的4×3维矩阵。
位置误差方程:
Figure PCTCN2019079136-appb-000087
Figure PCTCN2019079136-appb-000088
Figure PCTCN2019079136-appb-000089
其中,R M和R N分别表示地球的子午圈曲率半径和卯酉圈曲率半径,h表示载体所处的位置高度。
SINS***的噪声:
W N(t)=[ω gx ω gy ω gz ω ax ω ay ω az] T
式中,ω gi(i=x,y,z)表示三轴陀螺高斯白噪声,ω ai(i=x,y,z)表示三轴加速度计三轴高斯白噪声。
则SINS的***误差方程可以表示为:
Figure PCTCN2019079136-appb-000090
F N[·]为非线连续函数。
取DVL速度偏移误差(δV dx δV dy δV dz)和刻度系数误差(Δk dx Δk dy Δk dz)作为DVL***状态变量,记为:
X D(t)=[δV dx δV dy δV dz δk dx δk dy δk dz] T
DVL的误差模型为:
Figure PCTCN2019079136-appb-000091
Figure PCTCN2019079136-appb-000092
其中,β d表示速度偏移误差的相关时间倒数,ω d表示激励白噪声。
相应的误差状态方程为:
Figure PCTCN2019079136-appb-000093
式中:
Figure PCTCN2019079136-appb-000094
W D(t)=[ω dx ω dy ω dz] T;ω di(i=x,y,z)为激励白噪声;G D(t)=[I 3×3 O 3×3];τ di(i=x,y,z)表示速度偏移误差的相关时间。
选取SINS和DVL子***的误差状态变量,则组合导航***的状态向量为
Figure PCTCN2019079136-appb-000095
***的噪声向量为
Figure PCTCN2019079136-appb-000096
***的状态方程可表示为:
Figure PCTCN2019079136-appb-000097
式中,状态函数F 1[·]为非线性连续函数,Γ 1(t)为该子***噪声阵。
由于DVL测得地速在载体坐标系内的分量,要使其输出的速度与SINS输出的速度形成量测量,必须将DVL的输出速度变换到导航坐标系中。其中,
由SINS提供变化所用的姿态矩阵
Figure PCTCN2019079136-appb-000098
因此,由SINS和DVL形成的量测量
Figure PCTCN2019079136-appb-000099
将上式展开,并结合之前选取的***误差状态
Figure PCTCN2019079136-appb-000100
可得该组合***的量测方程为:
Figure PCTCN2019079136-appb-000101
式中:H 1[·]为非线性连续函数;V 1(t)为量测噪声。
据此,得到潜航阶段卡尔曼滤波模型的状态方程和量测方程:
Figure PCTCN2019079136-appb-000102
步骤三:选取状态量和量测量建立水面位置修正阶段的非线性***模型,取状态变量:
Figure PCTCN2019079136-appb-000103
***的噪声误差为:
W 2(t)=[0 0 ω gx ω gy ω gz ω ax ω ay 0 0 0 0 0] T
建立基于SINS/GPS的AUV组合导航连续***状态方程:
Figure PCTCN2019079136-appb-000104
式中:F 2[·]为非线性连续函数,Γ 2(t)为该子***噪声阵。
将SINS解算出来的位置、速度与GPS输出的位置位置、速度之差作为AUV水面位置修正阶段滤波解算的量测方程:
Figure PCTCN2019079136-appb-000105
式中L、λ、V E和V N分别为SINS解算出来的位置、速度,L G、λ G、V GE和V GN分别为GPS输出的位置位置、速度。将上式展开,并结合之前选取的***误差状态X 2(t),可得到量测方程为:
Z 2=H 2[X 2(t),t]+V 2(t)
式中:H 2[·]为非线性连续函数,V 2为量测噪声。
据此,得到水面位置修正阶段卡尔曼滤波模型的状态方程和量测方程为:
Figure PCTCN2019079136-appb-000106
步骤四:对上述建立的方程离散化,进行按照图4所示的算法改进的无迹卡尔曼滤波解算,实现时间更新、量测更新、滤波更新。
改进的卡尔曼滤波算法介绍如下:
针对l维状态变量,至少需要l+1样点来描述其均值和方差。球面分布单形采样变换(SSUT)就是通过l+1个分布在以状态均值为原点的球面上的权值相等的采样点来近似状态的概率分布。这样,l+1个球面分布的采样点和状态的均值点构成了l+2个无迹(UT)采样点。将这些点带入非线性函数中,相应的得到非线性函数数值点集,通过这些点集求取变换后的均值和协方差。
对于l维状态变量,球面分布单形采样变换(SSUT)采样点的具体选取步骤如下:
<1>选择0≤W 0≤1
<2>确定sigma点权值
W i=(1-W 0)/(L+1),i=1,2,...,l+1
<3>初始化向量序列
Figure PCTCN2019079136-appb-000107
<4>扩展向量序列(维数j=2,...l)
Figure PCTCN2019079136-appb-000108
其中,
Figure PCTCN2019079136-appb-000109
表示j维随机变量的第i个采样点,O j表示j维零向量。
<5>均值为
Figure PCTCN2019079136-appb-000110
均方差为P xx的L维随机变量x的球面分布点为:
Figure PCTCN2019079136-appb-000111
在无迹卡尔曼(UKF)推算法中,一般须对***噪声和量测噪声进行状态增广,但是当***噪声和量测噪声均为加性噪声时,可不做增广处理,有利于进一步降低滤波计算。本发明研究了一种基于复杂加性噪声的SSUT采样简化UKF算法。复杂加性噪声非线性离散***模型可表示为:
Figure PCTCN2019079136-appb-000112
式中:f[·]、g[·]、h[·]、j[·]均为非线性函数;x k、z k分别为状态向量和观测向量;ω k和v k分别为***状态噪声和量测噪声向量。其统计特性如下:E[W k]=0,
Figure PCTCN2019079136-appb-000113
E[V k]=0,
Figure PCTCN2019079136-appb-000114
由上式定义的***模型可知,复杂加性噪声模型的特点是模型关于噪声是线性的,具体的算法流程如下:
<1>初始化增广状态向量及估计误差方差
Figure PCTCN2019079136-appb-000115
Figure PCTCN2019079136-appb-000116
<2>计算sigma点和相应的加权因子
Figure PCTCN2019079136-appb-000117
W i=(1-W 0)/L+1,其中,0≤W 0≤1
<3>时间更新
χ i,k/k-1=f(χ i,k-1)
Figure PCTCN2019079136-appb-000118
Figure PCTCN2019079136-appb-000119
Z i,k/k-1=h(χ i,k/k-1)
Figure PCTCN2019079136-appb-000120
<4>量测更新
Figure PCTCN2019079136-appb-000121
Figure PCTCN2019079136-appb-000122
Figure PCTCN2019079136-appb-000123
Figure PCTCN2019079136-appb-000124
Figure PCTCN2019079136-appb-000125
根据上述算法,对上述建立的水下潜航阶段卡尔曼滤波方程进行离散化。
对于水下潜航阶段的滤波方程
Figure PCTCN2019079136-appb-000126
离散化有,
Figure PCTCN2019079136-appb-000127
其中,X k和Z k分别为***在t k时刻的状态向量以及量测向量,W k和V k分别为水下潜航阶段子***的噪声阵和量测噪声阵,且均值均为零,统计特性如下:
Figure PCTCN2019079136-appb-000128
Q k和R k分别为子***噪声协方差阵和量测噪声协方差阵。
具体的算法流程如下:
<1>初始化增广状态向量及估计误差方差
Figure PCTCN2019079136-appb-000129
Figure PCTCN2019079136-appb-000130
<2>计算sigma点和相应的加权因子
Figure PCTCN2019079136-appb-000131
W i=(1-W 0)/22,其中,0≤W 0≤1
其中,
Figure PCTCN2019079136-appb-000132
表示21维状态量的第i个采样点。
<3>时间更新方程,得到一步预测
Figure PCTCN2019079136-appb-000133
和一步预测误差协方差P k/k-1
χ i,k/k-1=f 1i,k-1)
Figure PCTCN2019079136-appb-000134
Figure PCTCN2019079136-appb-000135
<4>量测更新方程,得到k时刻的量测预测
Figure PCTCN2019079136-appb-000136
量测预测协方差
Figure PCTCN2019079136-appb-000137
和状态量与量测量之间的协方差
Figure PCTCN2019079136-appb-000138
Z i,k/k-1=h 1i,k/k-1)
Figure PCTCN2019079136-appb-000139
Figure PCTCN2019079136-appb-000140
Figure PCTCN2019079136-appb-000141
<5>滤波更新方程,得到滤波增益矩阵K k、状态量的最优滤波估计
Figure PCTCN2019079136-appb-000142
和最优滤波估计误差协方差矩阵P k
Figure PCTCN2019079136-appb-000143
Figure PCTCN2019079136-appb-000144
Figure PCTCN2019079136-appb-000145
对上述建立的水面位置修正阶段卡尔曼滤波方程进行离散化。对于水面位置修正阶段的滤波方程
Figure PCTCN2019079136-appb-000146
离散化有
Figure PCTCN2019079136-appb-000147
其中,x k和z k分别为***在t k时刻的状态向量以及量测向量,w k和v k分别为水面位置修正阶段子***的噪声阵和量测噪声阵,且均值均为零,,统计特性如下:
Figure PCTCN2019079136-appb-000148
Figure PCTCN2019079136-appb-000149
q k和r k分别为该子***噪声协方差阵和量测噪声协方差阵。具体算法流程如下:
<1>初始化增广状态向量及估计误差方差
Figure PCTCN2019079136-appb-000150
Figure PCTCN2019079136-appb-000151
<2>计算sigma点和相应的加权因子
Figure PCTCN2019079136-appb-000152
W i=(1-W 0)/13,其中,0≤W 0≤1
其中,
Figure PCTCN2019079136-appb-000153
表示12维状态量的第i个采样点。
<3>时间更新方程,得到一步预测
Figure PCTCN2019079136-appb-000154
和一步预测误差协方差P k/k-1
χ i,k/k-1=f 2i,k-1)
Figure PCTCN2019079136-appb-000155
Figure PCTCN2019079136-appb-000156
<4>量测更新方程,得到k时刻的量测预测
Figure PCTCN2019079136-appb-000157
量测预测协方差
Figure PCTCN2019079136-appb-000158
和状态量与量测量之间的协方差
Figure PCTCN2019079136-appb-000159
Z i,k/k-1=h 2i,k/k-1)
Figure PCTCN2019079136-appb-000160
Figure PCTCN2019079136-appb-000161
Figure PCTCN2019079136-appb-000162
<5>滤波更新方程,得到滤波增益矩阵K k、状态量的最优滤波估计
Figure PCTCN2019079136-appb-000163
和最优滤波估计误差协方差矩阵P k
Figure PCTCN2019079136-appb-000164
Figure PCTCN2019079136-appb-000165
Figure PCTCN2019079136-appb-000166
将每次AUV浮出水面的时的滤波结果的位置信息作为下一次潜航的新的位置信息,从而实现位置的定时修正,克服惯导的累积误差。
本发明方案所公开的技术手段不仅限于上述实施方式所公开的技术手段,还包括由以上技术特征任意组合所组成的技术方案。应当指出,对于本技术领域的普通技术人员来说,在不脱离本发明原理的前提下,还可以做出若干改进和润饰,这些改进和润饰也视为本发明的保护范围。

Claims (8)

  1. 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,包括如下步骤:
    (1)对需要用到的坐标系进行定义:
    (2)建立水下潜航阶段SINS/DVL子***的状态方程:以SINS的3轴姿态误差,3轴速度误差,位置误差,3轴陀螺仪随机漂移、3轴加速度计零偏和DVL速度偏移误差、刻度系数误差建立21维状态向量,根据***误差方程建立该子***的状态方程;
    (3)建立水下潜航阶段SINS/DVL子***的量测方程:根据SINS解算的3维速度量和DVL测量的速度量之差作为量测量,并结合步骤(2)选取的误差状态向量建立该子***的量测方程;
    (4)建立水面位置修正阶段SINS/GPS子***的状态方程:不考虑天向运动速度及误差,以SINS的3轴姿态误差、速度误差、位置误差、3轴陀螺仪随机漂移、加速度计零偏建立12维状态向量,根据误差方程建立该子***的状态方程;
    (5)建立水面位置修正阶段SINS/GPS子***的量测方程:以SINS解算出来的位置、速度与GPS输出的位置、速度之差作为量测量,并结合步骤(4)选取的误差状态向量建立该子***的量测方程;
    (6)综合步骤(2)和步骤(3)建立水下潜航阶段的非线性滤波方程,在水下潜航一段时间以后,AUV浮出水面,结合步骤(4)和步骤(5)建立水面位置修正阶段非线性滤波方程,进行改进的无迹卡尔曼滤波解算,完成时间更新、量测更新和滤波更新,完成定时位置信息校正。
  2. 根据权利要求1所述的改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,所述步骤(1)建立的坐标系具体包括:
    i——惯性坐标系:不随地球旋转,原点位于地球中心,z i轴指向北极,x i轴指向春分点,y i轴与x i、z i构成右手坐标系;
    e——地球坐标系:与地球固联,原点位于地心,x e轴穿越本初子午线与赤道交点,z e轴指向北极,y e轴x e、z e构成右手坐标系;
    b——载体坐标系:原点位于运载体中心,z b轴垂直运载体向上,x b指向运载体前方,y b与x b、z b构成右手坐标系;
    p——实际计算得出的平台坐标系;
    n——与东-北-天地理坐标系重合的导航坐标系。
  3. 根据权利要求1所述的改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,所述步骤(2)中建立水下潜航阶段SINS/DVL子***的状态方程具体包括如下步骤:
    取姿态误差角(φ E φ N φ U)、速度误差(δv E δv N δv U)位置误差(δL δλ δh)、陀螺常值漂移(ε bx ε by ε bz)和加速度计随机常值误差
    Figure PCTCN2019079136-appb-100001
    作为SINS***的状态量,记为:
    Figure PCTCN2019079136-appb-100002
    通过理想惯导比力方程的速度微分方程和捷联惯导***的实际速度微分方程,推导出四元数表示的速度误差方程:
    Figure PCTCN2019079136-appb-100003
    其中,Vn是载体在n系下的理想速度,
    Figure PCTCN2019079136-appb-100004
    是n系下的地球自转角速度,
    Figure PCTCN2019079136-appb-100005
    为n系相对于e系的角速度在n系下的投影,f b为b系下的比力;式中,
    Figure PCTCN2019079136-appb-100006
    为p系到n系的转换四元数,
    Figure PCTCN2019079136-appb-100007
    为b系到p系的转换四元数,
    Figure PCTCN2019079136-appb-100008
    Figure PCTCN2019079136-appb-100009
    分别代表p系到n系以及b系到p系的转换矩阵;
    Figure PCTCN2019079136-appb-100010
    式中,“~”表示载体实际测量值,δ表示载体的理想值与实际测量值之间的误差,g n为n系下的重力加速度,
    Figure PCTCN2019079136-appb-100011
    为b系下的加速度计误差向量;
    四元数姿态误差方程:
    Figure PCTCN2019079136-appb-100012
    其中,
    Figure PCTCN2019079136-appb-100013
    表示n系相对于i系的旋转角速度在n系下的投影,
    Figure PCTCN2019079136-appb-100014
    为n系到p系的方向余弦矩阵,ε b为陀螺仪误差向量在b系下的投影,B为关于四元数的4×3维矩阵;
    位置误差方程:
    Figure PCTCN2019079136-appb-100015
    Figure PCTCN2019079136-appb-100016
    Figure PCTCN2019079136-appb-100017
    其中,R M和R N分别表示地球的子午圈曲率半径和卯酉圈曲率半径,L表示载体的纬度,λ表示载体的经度,h表示载体的高度;
    SINS***的噪声:
    W N(t)=[ω gx ω gy ω gz ω ax ω ay ω az] T
    则SINS的***误差方程可以表示为:
    Figure PCTCN2019079136-appb-100018
    F N[·]为非线连续函数;
    取DVL速度偏移误差(δV dx δV dy δV dz)和刻度系数误差(Δk dx Δk dy Δk dz)作为DVL***状态变量,记为:
    X D(t)=[δV dx δV dy δV dz δk dx δk dy δk dz] T
    DVL的误差模型为:
    Figure PCTCN2019079136-appb-100019
    Figure PCTCN2019079136-appb-100020
    其中,β d表示速度偏移误差的相关时间倒数,ω d表示激励白噪声;
    相应的误差状态方程为:
    Figure PCTCN2019079136-appb-100021
    式中:
    Figure PCTCN2019079136-appb-100022
    W D(t)=[ω dx ω dy ω dz] T;ω di(i=x,y,z)为激励白噪声;G D(t)=[I 3×3 O 3×3];τ di(i=x,y,z)表示速度偏移误差的相关时间;
    选取SINS和DVL子***的误差状态变量,则组合导航***的状态向量为
    Figure PCTCN2019079136-appb-100023
    *** 的噪声向量为
    Figure PCTCN2019079136-appb-100024
    ***的状态方程表示为:
    Figure PCTCN2019079136-appb-100025
    式中,状态函数F 1[·]为非线性连续函数,Γ 1(t)为该子***噪声阵。
  4. 根据权利要求1所述的改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,所述步骤(3)中建立水下潜航阶段SINS/DVL子***的量测方程具体包括如下步骤:
    由SINS和DVL形成的量测量:
    Figure PCTCN2019079136-appb-100026
    其中,由SINS提供变化所用的姿态矩阵
    Figure PCTCN2019079136-appb-100027
    v EI,v NI,v UI分别表示SINS解算出的三轴速度,v ED,v ND,v UD分别表示DVL测出的三轴速度,v E,v N,v U表示载体在导航坐标系下的真实速度,v x,v y,v z表示在载体坐标系下的真实速度,δv DE,δv DN,δv DU为DVL测速误差换算到导航坐标系后的误差;
    将上式展开,并结合之前选取的***误差状态
    Figure PCTCN2019079136-appb-100028
    得该组合***的量测方程为:
    Z 1=H 1[X 1(t),t]+V 1(t)
    式中:H 1[·]为非线性连续函数;V 1(t)为量测噪声。
  5. 根据权利要求1所述的改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,所述步骤(4)中建立水面位置修正阶段SINS/GPS子***的状态方程具体包括如下步骤:
    取状态变量:
    Figure PCTCN2019079136-appb-100029
    ***的噪声误差为:
    W 2(t)=[0 0 ω gx ω gy ω gz ω ax ω ay 0 0 0 0 0] T
    建立基于SINS/GPS的AUV组合导航连续***状态方程:
    Figure PCTCN2019079136-appb-100030
    式中:F 2[·]为非线性连续函数,Γ 2(t)为该子***噪声阵。
  6. 根据权利要求1所述的改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,所述步骤(5)中建立水面位置修正阶段SINS/GPS子***的量测方程具体包括如下步骤:
    将SINS解算出来的位置、速度与GPS输出的位置位置、速度之差作为AUV水面位置修正阶段滤波解算的量测方程:
    Figure PCTCN2019079136-appb-100031
    式中L、λ、V E和V N分别为SINS解算出来的位置、速度,L G、λ G、V GE和V GN分别为GPS输出的位置位置、速度,δ·表示相应的误差;将上式展开,并结合之前选取的***误差状态X 2(t),得到量测方程为:
    Z 2=H 2[X 2(t),t]+V 2(t)
    式中:H 2[·]为非线性连续函数,V 2为量测噪声。
  7. 根据权利要求1所述的改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,所述步骤(6)中综合步骤(2)和步骤(3)建立的水下潜航阶段的非线性滤波方程为:
    Figure PCTCN2019079136-appb-100032
    结合步骤(4)和步骤(5)建立的水面位置修正阶段非线性滤波方程为:
    Figure PCTCN2019079136-appb-100033
  8. 根据权利要求7所述的改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法,其特征在于,所述步骤(6)中进行改进的无迹卡尔曼滤波解算的过程具体包括如下步骤:
    对水下潜航阶段的非线性滤波方程离散化有:
    Figure PCTCN2019079136-appb-100034
    其中,X k和Z k分别为***在t k时刻的状态向量以及量测向量,W k和V k分别为水下潜航阶段子***的噪声阵和量测噪声阵,且均值均为零,统计特性如下:
    Figure PCTCN2019079136-appb-100035
    Q k和R k分别为子***噪声协方差阵和量测噪声协方差阵;具体的算法过程如下:
    <1>初始化增广状态向量及估计误差方差
    Figure PCTCN2019079136-appb-100036
    Figure PCTCN2019079136-appb-100037
    <2>计算sigma点χ i,k-1和相应的加权因子W i
    Figure PCTCN2019079136-appb-100038
    W i=(1-W 0)/22,其中,0≤W 0≤1
    其中,
    Figure PCTCN2019079136-appb-100039
    表示21维状态量的第i个采样点;
    <3>时间更新方程,得到一步预测
    Figure PCTCN2019079136-appb-100040
    和一步预测误差协方差P k/k-1
    χ i,k/k-1=f 1i,k-1)
    Figure PCTCN2019079136-appb-100041
    Figure PCTCN2019079136-appb-100042
    χ i,k/k-1为t k时刻预测的第i个样本点;
    <4>量测更新方程,得到k时刻的量测预测
    Figure PCTCN2019079136-appb-100043
    量测预测协方差
    Figure PCTCN2019079136-appb-100044
    和状态量与量测量之间的协 方差
    Figure PCTCN2019079136-appb-100045
    Z i,k/k-1=h 1i,k/k-1)
    Figure PCTCN2019079136-appb-100046
    Figure PCTCN2019079136-appb-100047
    Figure PCTCN2019079136-appb-100048
    其中,Z i,k/k-1为第i个量测值;
    <5>滤波更新方程,得到滤波增益矩阵K k、状态量的最优滤波估计
    Figure PCTCN2019079136-appb-100049
    和最优滤波估计误差协方差矩阵P k
    Figure PCTCN2019079136-appb-100050
    Figure PCTCN2019079136-appb-100051
    Figure PCTCN2019079136-appb-100052
    对水面位置修正阶段非线性滤波方程进行离散化有:
    Figure PCTCN2019079136-appb-100053
    其中,x k和z k分别为***在t k时刻的状态向量以及量测向量,w k和v k分别为水面位置修正阶段子***的噪声阵和量测噪声阵,且均值均为零,统计特性如下:
    Figure PCTCN2019079136-appb-100054
    Figure PCTCN2019079136-appb-100055
    q k和r k分别为该子***噪声协方差阵和量测噪声协方差阵;
    具体算法步骤如下:
    <1>初始化增广状态向量及估计误差方差
    Figure PCTCN2019079136-appb-100056
    Figure PCTCN2019079136-appb-100057
    <2>计算sigma点χ i,k-1和相应的加权因子W i
    Figure PCTCN2019079136-appb-100058
    W i=(1-W 0)/13,其中,0≤W 0≤1
    其中,
    Figure PCTCN2019079136-appb-100059
    表示12维状态量的第i个采样点;
    <3>时间更新方程,得到一步预测
    Figure PCTCN2019079136-appb-100060
    和一步预测误差协方差P k/k-1
    χ i,k/k-1=f 2i,k-1)
    Figure PCTCN2019079136-appb-100061
    Figure PCTCN2019079136-appb-100062
    χ i,k/k-1为t k时刻预测的第i个样本点;
    <4>量测更新方程,得到k时刻的量测预测
    Figure PCTCN2019079136-appb-100063
    量测预测协方差
    Figure PCTCN2019079136-appb-100064
    和状态量与量测量之间的协方差
    Figure PCTCN2019079136-appb-100065
    Z i,k/k-1=h 2i,k/k-1)
    Figure PCTCN2019079136-appb-100066
    Figure PCTCN2019079136-appb-100067
    Figure PCTCN2019079136-appb-100068
    其中,Z i,k/k-1为第i个量测值;
    <5>滤波更新方程,得到滤波增益矩阵K k、状态量的最优滤波估计
    Figure PCTCN2019079136-appb-100069
    和最优滤波估计误差协方差矩阵P k
    Figure PCTCN2019079136-appb-100070
    Figure PCTCN2019079136-appb-100071
    Figure PCTCN2019079136-appb-100072
    将每次AUV浮出水面的时的滤波结果的位置信息作为下一次潜航的新的位置信息,定时修正位置。
PCT/CN2019/079136 2018-09-30 2019-03-21 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法 WO2020062807A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811159410.5 2018-09-30
CN201811159410.5A CN109141436A (zh) 2018-09-30 2018-09-30 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法

Publications (1)

Publication Number Publication Date
WO2020062807A1 true WO2020062807A1 (zh) 2020-04-02

Family

ID=64814207

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/079136 WO2020062807A1 (zh) 2018-09-30 2019-03-21 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法

Country Status (2)

Country Link
CN (1) CN109141436A (zh)
WO (1) WO2020062807A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111982117A (zh) * 2020-08-17 2020-11-24 电子科技大学 一种基于深度学习的auv光学引导与测向方法
CN112393729A (zh) * 2020-10-29 2021-02-23 国家深海基地管理中心 一种深海水面移动声学基准层联合导航定位***及方法
CN112432644A (zh) * 2020-11-11 2021-03-02 杭州电子科技大学 基于鲁棒自适应无迹卡尔曼滤波的无人艇组合导航方法
CN113503891A (zh) * 2021-04-22 2021-10-15 中国人民解放军海军工程大学 一种sinsdvl对准校正方法、***、介质及设备

Families Citing this family (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109141436A (zh) * 2018-09-30 2019-01-04 东南大学 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法
CN109931936A (zh) * 2019-03-18 2019-06-25 西北工业大学 一种基于移动中继站的弱连通auv协同导航方法
CN110926497B (zh) * 2019-04-08 2020-10-16 青岛中海潮科技有限公司 水下运载器惯导误差预测、浮起校正自动规划方法、装置
CN110208843B (zh) * 2019-05-21 2022-07-22 南京航空航天大学 一种基于增广伪距信息辅助的容错导航方法
CN110146076B (zh) * 2019-06-06 2023-04-18 哈尔滨工业大学(威海) 一种无逆矩阵自适应滤波的sins/dvl组合定位方法
CN110514203B (zh) * 2019-08-30 2022-06-28 东南大学 一种基于isr-ukf的水下组合导航方法
CN112230296B (zh) 2019-12-17 2021-07-23 东南大学 一种重力相关时间倒数确定方法
CN111366962A (zh) * 2020-03-12 2020-07-03 国家深海基地管理中心 一种深远海低成本长航时协同导航定位***
CN111722295B (zh) * 2020-07-04 2021-04-23 东南大学 一种水下捷联式重力测量数据处理方法
CN112254718B (zh) * 2020-08-04 2024-04-09 东南大学 一种运动约束辅助的基于改进Sage-Husa自适应滤波的水下组合导航方法
CN111982126B (zh) * 2020-08-31 2023-03-17 郑州轻工业大学 一种全源BeiDou/SINS弹性状态观测器模型设计方法
CN112684207B (zh) * 2020-12-17 2022-03-11 东南大学 一种深潜载人潜水器adcp速度估计与修正算法
CN112729291B (zh) * 2020-12-29 2022-03-04 东南大学 一种深潜长航潜水器sins/dvl洋流速度估计方法
CN113654559B (zh) * 2021-08-20 2023-03-31 青岛澎湃海洋探索技术有限公司 基于多模型观测校正的auv导航方法
CN113959433B (zh) * 2021-09-16 2023-12-08 南方电网数字平台科技(广东)有限公司 一种组合导航方法及装置
CN114777745A (zh) * 2022-04-08 2022-07-22 南京信息工程大学 一种基于无迹卡尔曼滤波的倾斜取证建模方法
CN117606491B (zh) * 2024-01-24 2024-04-26 中国海洋大学 一种自主式水下航行器的组合定位导航方法及装置

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103278163A (zh) * 2013-05-24 2013-09-04 哈尔滨工程大学 一种基于非线性模型的sins/dvl组合导航方法
CN103744098A (zh) * 2014-01-23 2014-04-23 东南大学 基于sins/dvl/gps的auv组合导航***
CN109141436A (zh) * 2018-09-30 2019-01-04 东南大学 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103278163A (zh) * 2013-05-24 2013-09-04 哈尔滨工程大学 一种基于非线性模型的sins/dvl组合导航方法
CN103744098A (zh) * 2014-01-23 2014-04-23 东南大学 基于sins/dvl/gps的auv组合导航***
CN109141436A (zh) * 2018-09-30 2019-01-04 东南大学 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
LIU, MINGYONG ET AL.: "Research on Integrated Navigation for Autonomous Underwater Vehicle based on an Improved Unscented Kalman Filter", ACTA ARMAMENTARII, vol. 32, no. 2, 28 February 2011 (2011-02-28), pages 252 - 255, XP055699666 *
LV, ZHIGANG: "Research on SINS/DVL/GPS Integrated Navigation System Calibration Method and the Error Analysis of AUV", SHIP ELECTRONIC ENGINEERING, vol. 38, no. 6, 30 June 2018 (2018-06-30), pages 33 - 36 *
ZHOU, BENCHUAN: "Federated filtering algorithm based on fuzzy adaptive UKF for marine SINS/GPS/DVL integrated system", 2010 CHINESE CONTROL AND DECISION CONFERENCE, 31 December 2010 (2010-12-31), XP031700090 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111982117A (zh) * 2020-08-17 2020-11-24 电子科技大学 一种基于深度学习的auv光学引导与测向方法
CN111982117B (zh) * 2020-08-17 2022-05-10 电子科技大学 一种基于深度学习的auv光学引导与测向方法
CN112393729A (zh) * 2020-10-29 2021-02-23 国家深海基地管理中心 一种深海水面移动声学基准层联合导航定位***及方法
CN112393729B (zh) * 2020-10-29 2021-07-16 国家深海基地管理中心 一种深海水面移动声学基准层联合导航定位***及方法
CN112432644A (zh) * 2020-11-11 2021-03-02 杭州电子科技大学 基于鲁棒自适应无迹卡尔曼滤波的无人艇组合导航方法
CN112432644B (zh) * 2020-11-11 2022-03-25 杭州电子科技大学 基于鲁棒自适应无迹卡尔曼滤波的无人艇组合导航方法
CN113503891A (zh) * 2021-04-22 2021-10-15 中国人民解放军海军工程大学 一种sinsdvl对准校正方法、***、介质及设备

Also Published As

Publication number Publication date
CN109141436A (zh) 2019-01-04

Similar Documents

Publication Publication Date Title
WO2020062807A1 (zh) 改进的无迹卡尔曼滤波算法在水下组合导航中的应用方法
CN109443379B (zh) 一种深海潜航器的sins/dvl水下抗晃动对准方法
CN108827310B (zh) 一种船用星敏感器辅助陀螺仪在线标定方法
CN110514203B (zh) 一种基于isr-ukf的水下组合导航方法
CN103759742B (zh) 基于模糊自适应控制技术的捷联惯导非线性对准方法
CN104655131B (zh) 基于istssrckf的惯性导航初始对准方法
CN106123921A (zh) 动态干扰条件下捷联惯导***的纬度未知自对准方法
Arnold et al. Robust model-aided inertial localization for autonomous underwater vehicles
CN101920762A (zh) 一种基于噪声矩阵实时修正的船舶动力定位方法
CN115200574A (zh) 一种地球椭球模型下的极区横向组合导航方法
Huang et al. Weight self-adjustment Adams implicit filtering algorithm for attitude estimation applied to underwater gliders
CN112015086B (zh) 一种欠驱动水面船有限时间路径跟踪输出反馈控制方法
Ben et al. A dual-state filter for a relative velocity aiding strapdown inertial navigation system
Xu et al. USBL positioning system based adaptive kalman filter in AUV
CN115855049A (zh) 基于粒子群优化鲁棒滤波的sins/dvl导航方法
CN111964675A (zh) 一种黑障区的飞行器智能导航方法
CN104061930A (zh) 基于捷联惯性制导和多普勒计程仪的导航方法
CN110873577B (zh) 一种水下快速动基座对准方法及装置
Wang et al. AUV navigation based on inertial navigation and acoustic positioning systems
CN103616026B (zh) 一种基于h∞滤波的auv操纵模型辅助捷联惯导组合导航方法
CN105300407A (zh) 一种用于单轴调制激光陀螺惯导***的海上动态启动方法
CN103697887B (zh) 一种基于捷联惯导***和多普勒计程仪的优化导航方法
Zhang Autonomous underwater vehicle navigation using an adaptive Kalman filter for sensor fusion
Braga et al. Navigation scheme for the LSTS SEACON vehicles: Theory and application
Bao et al. AUV Docking Recovery Based on USBL Integrated Navigation Method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19865580

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19865580

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 19865580

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205N DATED 26/05/2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19865580

Country of ref document: EP

Kind code of ref document: A1