Mars landing section autonomous navigation method and device fusing point and line features

Abstract

The invention provides a Mars landing section autonomous navigation method and device fusing point and line features. The method comprises the following steps: acquiring measurement information of the surface of Mars; analyzing point features and line features in the measurement information, and constructing a measurement model; constructing a state model according to the measurement information and a dynamical model; establishing an autonomous navigation model by using the measurement model and the state model; and updating the navigation state through extended Kalman filtering according to the autonomous navigation model. The advantages and disadvantages of the point features and the line features in different scenes are analyzed, the fact that the line features have high robustness compared with the point features in the low-texture and complex environment is analyzed, and a nonlinear error with higher precision is introduced to define the measurement error of the line features. A digital simulation result shows that compared with a traditional measurement model which only uses point feature measurement, the method has higher navigation precision.

Description

Technical Field

[0001] The present invention relates to the technical field of navigation observation, and particularly to an autonomous navigation method and device for the Mars landing section that integrates point and line features. Background Art

[0002] Autonomous and precise landing is the key to the Mars exploration mission. The primary prerequisite is that the spacecraft can autonomously obtain high-precision state information such as position, velocity, and attitude, that is, autonomous navigation, so as to move to the predetermined landing point. Due to the long time and long distance of the Mars exploration mission and the satellite navigation denial, and since the error of inertial navigation accumulates with time and can only maintain a high precision within a short time, in order to achieve high-precision autonomous navigation throughout the process, it is necessary to accurately estimate and compensate the accumulated error of inertial navigation during the powered descent process. During the Mars orbiter's exploration of Mars, the Mars surface image information is obtained by high-frequency shooting of the on-board camera, which provides an effective measurement information reference for the autonomous navigation in the landing section. During the landing process, the on-board camera takes pictures of the Mars surface to extract landmark information and matches it with the known landmark information to obtain navigation information. Currently, landmark-based integrated navigation has become a popular research direction for autonomous navigation in the landing section. Most traditional landmark-based integrated navigation methods use point features as measurement information. However, during the Mars landing process, as the altitude of the detector decreases, the observable land points become fewer and the measurement information is insufficient, resulting in a decrease in navigation accuracy.

[0003] In summary, the following problems exist in the prior art: the navigation accuracy of autonomous navigation in the Mars landing section is insufficient. Summary of the Invention

[0004] The purpose of the present invention is to solve the problem of how to improve the navigation accuracy of autonomous navigation in the Mars landing section.

[0005] To this end, on the one hand, an embodiment of the present invention provides an autonomous navigation method for the Mars landing section that integrates point and line features, and the method includes the following steps:

[0006] Obtain the measurement information of the Mars surface;

[0007] Analyze the point features and line features in the measurement information and construct a measurement model;

[0008] Construct a state model according to the measurement information and the dynamic model;

[0009] Establish an autonomous navigation model by using the measurement model and the state model;

[0010] Update the navigation state through extended Kalman filtering according to the autonomous navigation model.

[0011] On the other hand, an embodiment of the present invention provides an autonomous navigation device for the Mars landing section that integrates point and line features, including:

[0012] A measurement unit for obtaining measurement information of the Martian surface;

[0013] An analysis unit for analyzing point features and line features in the measurement information and constructing a measurement model;

[0014] A construction unit for constructing a state model based on the measurement information and a dynamics model;

[0015] A fusion unit for establishing an autonomous navigation model by using the measurement model and the state model;

[0016] A navigation unit for updating the navigation state through extended Kalman filtering according to the autonomous navigation model.

[0017] The above technical solution has the following beneficial effects: The present invention uses line features that are more robust than point features in low-texture scenarios to make up for the problem of insufficient measurement information, derives a measurement equation based on line feature measurement for the Mars landing section, and fuses the measurement equation of traditional point features. Combining with the state equation constructed based on inertial dynamics, it linearizes them by using extended Kalman filtering to correct the cumulative error of inertial navigation and achieve the optimal estimation of the system state. The results of mathematical simulation show that the measurement model that fuses point and line features has higher navigation accuracy than the traditional measurement using only point features. Description of the Drawings

[0018] Figure 1 is a flowchart of an autonomous navigation method for the Mars landing section that fuses point and line features provided by an embodiment of the present invention;

[0019] Figure 2 is a schematic structural diagram of an autonomous navigation device for the Mars landing section that fuses point and line features provided by an embodiment of the present invention;

[0020] Figure 3 is a schematic diagram of a reference coordinate system provided by an embodiment of the present invention;

[0021] Figure 4 is an image of the Martian surface;

[0022] Figure 5 is a schematic diagram of the matching error of the center point of an ellipse provided by an embodiment of the present invention;

[0023] Figure 6 is a geometric schematic diagram of the PNP problem provided by an embodiment of the present invention;

[0024] Figure 7 is a straight line extraction and reprojection error diagram provided by an embodiment of the present invention;

[0025] Figure 8 is a general framework diagram of the navigation system design provided by an embodiment of the present invention;

[0026] Figure 9 It is the point-line feature measurement model and error ellipsoid diagram provided by the embodiments of the present invention. Detailed implementation manners

[0027] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0028] The technical solution of the present invention is: a method for autonomous navigation in the Mars landing stage that fuses point and line features is proposed. The advantages and disadvantages of point features and line features in different scenarios are analyzed. Theoretically, it is analyzed that line features have higher robustness than point features in low-texture and complex environments. A more accurate non-linear error is introduced to define the measurement error of line features. Finally, the extended Kalman filter is used to linearize the measurement equation that fuses point and line features and combine it with the measurement equation established by inertial dynamics to achieve state update and estimation. The digital simulation results show that compared with the traditional measurement model that only uses point features for measurement, the measurement model that fuses point and line features can better solve the problem of insufficient measurement information in the latter stage of Mars landing and has higher navigation accuracy.

[0029] In the embodiments of the present invention, as Figure 1 , a method for autonomous navigation in the Mars landing stage that fuses point and line features is provided. The method includes the following steps:

[0030] S101: Obtain the measurement information on the Mars surface;

[0031] Measurement information on the Mars surface:

[0032] The soft landing of the detector on Mars is an important part of Mars exploration. It is necessary to determine the position, attitude and velocity of the detector relative to the selected reference system. For ease of description, the following 3 coordinate systems are introduced: the landing coordinate system, the vehicle coordinate system, and the camera coordinate system, as Figure 3 shown.

[0033] The landing coordinate system (L system) is centered on a predetermined landing point. The z-axis is perpendicular to the landing plane and points upward. The y-axis located in the landing plane points in the planetary rotation direction. The x L axis, y L axis and z L axis together construct a right-handed coordinate system.

[0034] The vehicle coordinate system (B system) The vehicle coordinate system is a coordinate system fixedly connected to the detector body. Its coordinate origin is the center of mass of the detector. The x BAxis, y B Axis and z B Axes are usually defined on the principal axes of inertia of the detector.

[0035] The camera coordinate system (C system) has the optical center of the camera as the origin of the coordinate system. The x C y C axes are parallel to the x and y axes of the image coordinate system, and the optical axis of the camera is the z C axis. The coordinate system satisfies the right-hand rule.

[0036] Point feature measurement:

[0037] The common landmark features on the Martian surface are generally composed of craters and impact craters. During the landing process, the Mars probe takes images of the Martian surface through an on-board camera, extracts the landmark points in the images, and thus obtains the navigation status information. The image of the Martian surface is as Figure 4 shown. However, the common craters and impact craters on the Martian surface cannot be regarded as a single landmark point. In most cases, it is an elliptical surface. For the sake of simplicity in processing, the center point of the ellipse is generally used as the landmark point for point feature extraction.

[0038] Using the center point of the landmark feature as the landmark point for smaller craters results in a smaller measurement error. However, for larger craters, continuing to use the above method will result in a larger measurement error. The analysis is as follows. Assume that there are two ellipses in the image taken by the on-board camera, with ellipse areas S1 and S2 respectively, where S1 >> S2. The coordinates of the center points of the ellipses in the landmark library are (x1, y1) and (x2, y2) respectively. Assume that there is no consideration of matching error, imaging error, etc., and only consider the deviation ξ i , and let's assume

[0039] The center point of the ellipse extracted from the image taken by the on-board camera can be expressed as:

[0040]

[0041] ( C x i , C y i ) are the coordinates of the extracted center point, (x i , y i ) are the true center point coordinates. The extracted center point coordinates will fall within a circle with the true coordinates as the center and a radius of ρ i , as Figure 5 shown.

[0042] The radius ρ i of the central circle is proportional to the area of the ellipse, and the variance σ i of the deviation ξ iProportional to the area of the ellipse, i.e., σ i ∝ρ i ∝S i . Calculate the residual between the center point of the ellipse extracted by the spaceborne camera and the true coordinates as

[0043]

[0044] The covariance matrix of the residual vector is

[0045]

[0046] The trace of the covariance matrix represents the distance from the sample point to the sample center. In this paper, the trace (tr) of the covariance matrix is used to measure the dispersion degree of the sample points. That is, the larger the area of the ellipse, the greater the measurement error generated in the process of extracting the center point of the ellipse. Therefore, there is a large error in using the traditional point feature measurement method for large - area meteorite craters, and more robust measurement features need to be introduced.

[0047] Line feature measurement

[0048] For any two points p f1 and p f2 on a given straight line in the space rectangular coordinate system, the Plucker coordinates can be expressed as:

[0049]

[0050] The Plucker coordinate representation is similar to the coordinate transformation in the rectangular coordinate system when representing coordinate transformation and has linearity. Let the rotation and translation corresponding to the world coordinate system to the camera coordinate system be R c , t c . For the three - dimensional straight line l w in the world coordinate system and the three - dimensional straight line l c in the camera coordinate system, its coordinate transformation formula is

[0051]

[0052] Let C p f1 and C p f2 be the points of p f1 and p f2 in the pixel coordinate system, and convert them to homogeneous coordinates (u f1 , v f1 , 1), (u f2 , v f2 , 1), then there is C p f1 = Kp f1 , C pf2 = Kp f2 , Then the equation of the projected line of the space line in the camera coordinate system is

[0053]

[0054] where K is the matrix composed of the internal parameters of the camera. The above formula shows that the line projection of the line is only related to the normal vector and has nothing to do with the direction vector.

[0055] Currently, there are two definitions of the error between the space line and the matching line, the 3D-3D method and the 3D-2D method. The 3D-2D definition method adopts a simple projection line measurement model, with small data storage and higher precision. The 3D-2D method first reprojects the space line onto the image, calculates the error between the projected line l and the matching line l c in the image, and records any two points on the line as p f1 and p f2 . Bartoli gave several definitions of error, such as a linear error e = l × l without the meaning of geometric error c , and the linear error defined by the algebraic distance and analyzed the time efficiency and optimized precision of various errors. This paper adopts a non-linear error definition method with higher precision

[0056]

[0057] Let the landmark coordinates containing measurement errors be C′ p f (u1′, v1′, 1), and has the following relationship with the true coordinates C p f =(u1, v1, 1)

[0058]

[0059] Assume that the deviation ξ i obeys a Gaussian distribution with a mean of 0, that is Calculate the residual vector:

[0060]

[0061] Calculate the covariance matrix of the residual vector

[0062]

[0063] Calculating the trace of the covariance matrix gives Extracting line features requires at least two points. If two point features are extracted simultaneously, the trace of the covariance matrix is twice that of extracting a single feature point, that is Since In the case of a larger elliptical area, the line feature has higher precision than the point feature. In summary, the measurement of line features based on Plücker has higher precision and robustness than traditional point feature measurement.

[0064] S102: Analyze the point features and line features in the measurement information and construct a measurement model; specifically including:

[0065] Perform pose estimation of point features;

[0066] Perform pose estimation of line features;

[0067] Construct a measurement model through pose estimation of point features and pose estimation of line features.

[0068] The measurement model is specifically:

[0069]

[0070] Where, is the measurement of the landmark p k at time t by the camera; f is the focal length of the camera; j is the noise-free measurement value; is the measurement noise, and its covariance matrix is ; is the vector from the spaceborne camera to the landmark.

[0071] S103: Construct a state model according to the measurement information and the dynamic model;

[0072] S104: Establish an autonomous navigation model using the measurement model and the state model; The autonomous navigation model is specifically:

[0073]

[0074] Where, r k is a vector block that fuses point-line measurement information; z imu =(a imu , ω imu ) is the inertial measurement unit of the IMU, and t k is the k-th moment, in seconds; is the error state vector at the k-th moment.

[0075] S105: Update the navigation state through extended Kalman filtering according to the autonomous navigation model.

[0076] The update of the navigation state through extended Kalman filtering according to the autonomous navigation model includes:

[0077] Linearize the measurement model,

[0078] Linearize the state model;

[0079] Calculate the propagation of the navigation state error covariance;

[0080] Through the extended Kalman filter algorithm, combined with the linearized measurement model and the linearized state model, update the navigation state and state covariance to achieve the optimal estimation of the spacecraft state, thereby completing autonomous navigation.

[0081] The pose estimation method based on point-line features specifically includes:

[0082] During the process of the on-board camera shooting, by comparing the captured image with the landmark library, the attitude and position of the carrier relative to the ground can be obtained. The process of determining the attitude and position can be obtained through modeling and calculation.

[0083] Point feature pose estimation includes:

[0084] The pose estimation based on point features is essentially a PNP problem (Perspective-N-Points problem). The PNP problem is a classical problem in computer vision, photogrammetry, and photogrammetry to study the geometric relationship of spatial perspective points. Given the known internal parameters and disparity coefficients of the camera, an image of N target points in space is captured, the spatial coordinates of the N target points are known, and the image coordinates of each target point are extracted through image analysis. Solve the relative position and attitude relationship between the camera and the spatial coordinate system defining these target points. The geometric schematic diagram is as Figure 6 shown.

[0085] The minimum configuration solution of the PNP problem is 6, that is, at least 6 feature points need to be observed to linearly solve the unique solution. The solution process is as follows. According to the conversion relationship between the landing system and the camera system, we can obtain

[0086]

[0087] where [X L , Y L , Z L are the coordinates of the feature points in the landing system, where are the coordinates of the feature points in the landing system, r0 to r8 are the elements of the rotation matrix from the landing system to the camera system, T X , T Y , T Z is the translation vector. Rewrite equation (11) as a homogeneous linear equation system about r0 to r8 and T X , T Y , T Z of

[0088]

[0089] T Z is related to the distance apart, then there is T Z > 0, let

[0090]

[0091] Rewrite (12) as a non - homogeneous linear equation system about s0~s 10 of the non - homogeneous linear equation system

[0092]

[0093] More than 6 non - coplanar feature points can be used to solve s simultaneously i , since the rotation matrix R is an orthonormal matrix, then there is

[0094]

[0095] The rotation angle can be decomposed to obtain

[0096]

[0097] If the feature points are coplanar, the minimum configuration solution of the PNP problem at this time is 4, that is, at least 4 feature points need to be observed to linearly solve the unique solution. The solution process is as follows. If the feature points are coplanar, then Z≡0, and equation (11) can be rewritten as

[0098]

[0099] Let

[0100]

[0101] Rewrite (18) as a linear equation system about s i (i = 0,1,3,4,6,7,9,10) of the linear equation system

[0102]

[0103] More than 4 coplanar feature points can be used to solve s simultaneously i , since the rotation matrix R is an orthonormal matrix, then there is

[0104]

[0105] The rotation angle can be decomposed to obtain

[0106]

[0107] Pose estimation of line features:

[0108] Pose estimation based on point features is essentially a PNL problem (Perspective-N-Lines problem). The minimum configuration solution for the PNL problem is 3. In a low-texture scene, currently, objects are relatively far away or the imaging is unclear, making it difficult to extract point features. Line features can reflect the geometric relationship of the object's position and are related to the position and pose of the target relative to it. According to the target pose parameters and the 3D model, reprojection is performed, which should be consistent with the actual image of the target. Taking the minimum deviation between the reprojection edge and the actual imaging edge as the objective function, the correction amount of the target pose parameters is calculated and the initial value is corrected, and the iteration continues until convergence, as Figure 7 shown

[0109] For the three-dimensional line l w in the landing system and the three-dimensional line l c in the camera system, their coordinate transformation formula is

[0110]

[0111] R c ,t c are the corresponding rotation matrix and translation vector. The corresponding relationship between the space line l w and the imaging line l c is l c =E·l w . Extract the target edge line on the image. Select the visible lines according to the initial pose and the target structure, perform reprojection, select several reprojection lines, and search for the nearest actual imaging line along the normal direction of the reprojection line. The distance d i . The actual pose differs from the initial value E by ΔE, and the actual image point is l c =E·ΔE·l w . Minimize the objective function to solve for ΔE and correct the target pose, and repeat the iteration until convergence.

[0112] Mars landing section system model

[0113] Dynamics model

[0114] Aiming at the problem that it is difficult to obtain an accurate planetary landing mechanics model with traditional dynamics models, this paper directly uses IMU (inertial navigation) measurement data to establish a dynamics model. The landing dynamics equation of the detector based on the IMU output is given. Assuming that the detector is affected by gravity and thrust, the landing dynamics model of the detector can be described as:

[0115]

[0116] where L v and L r are the position and velocity relative to the landing coordinate system (L system); is the quaternion from the L system to the B system, ba and b ω are the drift biases of the accelerometer and gyroscope; L a thrust is the control acceleration vector; L g is the gravitational action, calculated as a constant; n g is the unmodeled gravitational acceleration; n wa and n wω are the drift bias noises; ω = [ω x ω y ω z is the angular velocity of the object's rotation in the B frame. Ω(ω) is defined as:

[0117]

[0118] where is the skew-symmetric matrix of ω = [ω x ω y ω z , defined as

[0119]

[0120] The output values a imu and ω imu of the IMU can be expressed as

[0121]

[0122] where C q is the attitude rotation matrix corresponding to the quaternion q, L ω G is the angular velocity of the planet's rotation in the L frame, n a and n ω are the drift noises of the acceleration and angular velocity, is a zero-mean Gaussian white noise with covariance. Q is the covariance matrix,

[0123]

[0124] The state vector of the spacecraft is defined as:

[0125]

[0126] where L v and L r are the position and velocity in the relative landing coordinate system (L frame), is the quaternion from the L frame to the B frame, b a and b ω are the drift biases of the accelerometer and gyroscope. For the position, velocity, and biases, define an estimated value of a quantity a The error of is For quaternions, the quaternion error is described by δθ, corresponding to δq ∼ [δθ T 1].

[0127] Accordingly, the error state vector is defined as

[0128]

[0129] Combined with equation (25), the state equation for point-line feature-based navigation can be written as:

[0130]

[0131] Measurement model based on point-line features:

[0132] Using the pinhole projection model, at time t k of the camera, the measurement of landmark p j is

[0133]

[0134] where f is the camera focal length, is the noise-free measurement value, is the measurement noise, and its covariance matrix is is the vector from the spaceborne camera to the landmark, and the calculation formula is as follows

[0135]

[0136] L p j is the coordinate of the landmark in the L system and can be obtained by comparing the captured image with the landmark database in the landing map, represents the map error, and its covariance is L r k represents the coordinate of the detector in the L system. The geometric relationship is as Figure 8 shown.

[0137] Calculate the expected measurement value j of landmark p

[0138]

[0139] where,

[0140]

[0141] The observation residual is calculated as:

[0142]

[0143] At time t, the camera k has a reprojection error for the i-th land mark line as

[0144]

[0145] where s l =[u s , v s , 1], e l =[u e , v e , 1] are the homogeneous pixel coordinates of the two end points matching the line feature, and their coordinates in the camera system are respectively which can be obtained through (34), and the geometric relationship is as Figure 9 shown. This reprojection error projects the line feature from the three-dimensional space to the pixel coordinate system at time t k to obtain its representation in the pixel coordinate system Calculate the distance between the two end points of the line feature that matches it. Based on the measurement model of the line feature, it can be written as

[0146] Observability constraint:

[0147] Since the prior information of the land mark has been obtained before the actual flight, the true trajectory is unknown at the moment when the true land mark feature is observed. Whether the land mark feature is visible at time t depends on the true position r(t) and attitude q(t) of the detector at time t. The covariance matrix of the true position and attitude can be obtained at each time t.

[0148]

[0149] It shows that the true value r(t) of the position exists in the error ellipsoid centered at the estimated value with a certain probability, as Figure 9 shown. The length of the major axis of the 3σ error ellipsoid is where λ t,i is the eigenvalue of . The ellipsoid determined by is centered at the estimated value and is within the sphere with radius is the maximum eigenvalue of Ld, the maximum camera pointing error caused by attitude error is It can be obtained through calculation. For a landmark line feature, in the landing system, a point L r s on the spherical surface S to the landmark L p j vector can be written as:

[0150]

[0151] where L r l is a unit vector in any direction. For the landmark point feature L p j , if the following constraint is satisfied, the landmark point L p j is considered visible.

[0152]

[0153] Autonomous navigation method for fusing point and line measurement features

[0154] Combining (25)(37)(38), the autonomous navigation model for fusing point and line features can be written as

[0155]

[0156] where r k is a vector block that fuses point and line measurement information

[0157]

[0158] where j and i are the number of point and line features observed by the spacecraft at time t k respectively, and z imu =(a imu , ω imu ) is the inertial measurement unit of the IMU. Linearize the state equation

[0159]

[0160] The navigation state error covariance can be propagated through the numerical integration of the following formula

[0161]

[0162] where and G are Jacobian matrices, defined as follows

[29] , where

[0163]

[0164] The measurement model based on point-line features can be linearized as

[0165]

[0166] The measurement matrix based on point features is defined as

[30] :

[0167]

[0168] where

[0169]

[0170] The measurement matrix based on line features is defined as follows. At time t k of the camera, the measurement of the terrestrial reference line is where L p f1 , L p f2 are any two points on the terrestrial reference line. For the reprojection error of the line feature, the following partial derivatives can be calculated

[0171]

[0172] where

[0173]

[0174] For the navigation system described in (42), the extended Kalman filter (EKF) algorithm can be used to update the navigation state and state covariance as follows

[0175]

[0176] where diag(·) represents a matrix composed of diagonal components. Here, K(t k ) is the gain matrix, is the measurement innovation, and R j is the measurement noise covariance matrix combined with the map error.

[0177] Simulation

[0178] To evaluate the point-line fusion navigation scheme proposed in this paper, numerical simulation is carried out using MATLAB.

[0179] Simulation parameter settings:

[0180] The landing process on Mars was simulated, assuming that the duration of the landing phase was 100 s. The navigation camera recorded images up to 50 m at a rate of 1 Hz. The field of view (FOV) of this camera was 50°, the focal length was 1000 pixels, and the image noise was 1 pixel. Assume that the inertial navigation (IMU) is LN-200, providing inertial measurement values at 50 Hz. The nominal trajectory in the landing section was determined based on the Apollo Mars descent guidance algorithm. The true state of the detector at the initial moment is shown in Table 1. Assume that the initial position state error is 100 m, the initial velocity state error is 1 m / s, the initial attitude state error for each axis is 1°, and the rotational angular velocity is [0, 0, 0.2] ° / s. Assume that the distance from the landing point to the local horizontal plane follows a normal distribution with a standard deviation of 2 m, generating an irregular landing surface. Assume that the landing marks are randomly distributed on the irregular surface, and the standard deviation of the horizontal position error is 2 m. The process noise is Q. The local gravity noise is assumed to be Gaussian white noise with a standard deviation of 1×10 -1 m / s 2 of Gaussian white noise.

[0181] Q = diag([2.4×10 -14 I3 2.5×10 -14 I3 2.4×10 -7 I3 1.2×10 -7 I3 1×10 -8 I3])

[0182] Table 1 True state of the detector at the initial moment

[0183] State component True value <![CDATA[r0]]> <![CDATA[[1300 870 2570] T (m)]]> <![CDATA[v0]]> <![CDATA[[-28 -10 -90] T (m / s)]]> <![CDATA[q0]]> <![CDATA[[0.969 0.247 0 0] T > <![CDATA[b α > <![CDATA[[3×10 -3 3×10 -3 3×10 -3 T (m / s 2 )]]> ​ <![CDATA[b ω > <![CDATA[[4.8×10 -6 4.8×10 -6 4.8×10 -6 T (rad / s)]]> ​

[0184] Results of the simulation experiment:

[0185] The simulation recorded the nominal trajectory and all available landing points and landing lines, and gave the projection of the camera FOV along the nominal trajectory on the landing surface at each imaging moment. There are approximately 450 landing point features and 350 landing line features in the landing area. If only landing point features are used for navigation, in the middle and late stages of landing, when landing points cannot be observed or the landing points are blurred and invisible and there are fewer available landing points, the landing lines, as a robust measurement feature, are a good supplement to the landing points and can effectively improve the navigation accuracy.

[0186] Traditional navigation methods based on point feature measurement can greatly correct the cumulative error of inertial navigation and improve navigation accuracy. However, as the number of landmarks decreases during the landing process, the observable information is insufficient, and point feature-based measurement cannot meet the landing accuracy required by the spacecraft. To improve the navigation performance of landmark feature-based measurement, line feature measurement is introduced. The results show that the error curves of state estimation can converge to the ±3σ uncertainty bounds, and the measurement model that fuses point and line features can accurately estimate the state information of the spacecraft, with high navigation accuracy.

[0187] The root mean square error (RMSE) of position, velocity, and attitude using only point feature measurement and point-line fusion measurement in two different measurement methods. In the two different measurement modes of fusing point and line features and only using point features, during each landing stage, for the errors of position and attitude on each axis, the error of fusing point and line features is lower than the navigation error of only using point features. In the later stage of landing, as the number of observable landmark features decreases, the error also increases. However, for the measurement model that fuses point and line features, as the number of land points decreases in the later stage of landing, the land line features can be used as a supplement to point features, and the two are fused for estimation. Although the navigation error increases, it is still at a relatively low level compared with only using point feature measurement. The error is always within a tolerable range. For the errors and RMSE of velocity on each axis, at the beginning of landing, the velocity errors of both the method of fusing point and line features and only using point features are maintained at a relatively high level. As time goes by, the land mark measurement information corrects the inertial navigation measurement information multiple times, and the velocity error begins to decrease significantly. Compared with the land mark velocity error of only using point features, the line features, as a supplement to point features, can obtain more measurement information and have a better inhibitory effect on the error, and the error convergence speed is faster. In the later stage of landing, due to the correction of the cumulative error of the inertial navigation error by the measurement information, the errors of both are at a relatively low level and there is no significant difference. However, the velocity error of fusing point and line features is still lower than the velocity error of only using point features. Since the measurement information of the landmark does not directly include the measurement information of velocity, the correction of the velocity error is an indirect correction. The estimation of the detector velocity is mainly obtained by using the dynamic model, the measurement information is underdetermined, and the state estimation effect is unstable. There is no significant difference in the errors between the two in the absence of land mark measurement information.

[0188] In the embodiment of the present invention, as Figure 2 , a Mars landing section autonomous navigation device that fuses point and line features is also provided, including:

[0189] A measurement unit 21 for obtaining measurement information on the Mars surface;

[0190] An analysis unit 22 for analyzing the point features and line features in the measurement information and constructing a measurement model;

[0191] A construction unit 23 for constructing a state model according to the measurement information and the kinetic model;

[0192] A fusion unit 24 for establishing an autonomous navigation model by using the measurement model and the state model;

[0193] A navigation unit 25 for updating the navigation state through extended Kalman filtering according to the autonomous navigation model.

[0194] The analysis unit 22 includes:

[0195] Performing pose estimation of point features;

[0196] Performing pose estimation of line features;

[0197] Constructing a measurement model through pose estimation of point features and pose estimation of line features.

[0198] The measurement model is specifically:

[0199]

[0200] Where is the measurement of the landmark p by the camera at time t k ; f is the focal length of the camera; j is the noise-free measurement value; is the measurement noise, and its covariance matrix is is the vector from the spaceborne camera to the landmark.

[0201] The autonomous navigation model is specifically:

[0202]

[0203] Where r k is a vector block that fuses point and line measurement information; z imu =(a imu , ω imu ) is the inertial measurement unit of the IMU, and t k is the k-th moment, in seconds; is the error state vector at the k-th moment.

[0204] The navigation unit 25 includes:

[0205] Linearizing the measurement model,

[0206] Linearizing the state model;

[0207] Calculating the propagation of the navigation state error covariance;

[0208] Through the extended Kalman filter algorithm, combined with the linearized measurement model and the linearized state model, the navigation state and state covariance are updated to achieve the optimal estimation of the spacecraft state, thereby completing autonomous navigation.

[0209] The working principle of the autonomous navigation device for the Mars landing section that fuses point and line features is the same as that of the autonomous navigation method for the Mars landing section that fuses point and line features, and will not be elaborated here.

[0210] The present invention uses line features that are more robust than point features in low-texture scenarios to make up for the lack of measurement information, derives the measurement equation based on line feature measurement in the Mars landing section, and fuses the measurement equation of traditional point features. Combining with the state equation constructed based on inertial dynamics, the extended Kalman filter is used to linearize it to correct the cumulative error of inertial navigation and achieve the optimal estimation of the system state. The mathematical simulation results show that the measurement model that fuses point and line features has higher navigation accuracy than the traditional measurement using only point features.

[0211] It should be understood that the specific order or hierarchy of steps in the disclosed process is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process can be rearranged without departing from the scope of the present disclosure. The appended method claims present the elements of the various steps in an exemplary order and are not intended to be limited to the specific order or hierarchy recited.

[0212] In the above detailed description, various features are combined in a single embodiment to simplify the present disclosure. This method of disclosure should not be interpreted as reflecting an intention that the embodiments of the claimed subject matter require more features than are clearly recited in each claim. On the contrary, as reflected in the appended claims, the present invention resides in less than all of the features of a single disclosed embodiment. Accordingly, the appended claims are hereby expressly incorporated into the detailed description, with each claim standing alone as a separate preferred embodiment of the present invention.

[0213] In order to enable any person skilled in the art to implement or use the present invention, the above-described disclosed embodiments have been described. For those skilled in the art, various modification methods of these embodiments are obvious, and the general principles defined herein can also be applied to other embodiments without departing from the spirit and scope of the present disclosure. Therefore, the present disclosure is not limited to the embodiments given herein, but is consistent with the broadest scope of the principles and novel features disclosed in this application.

[0214] The foregoing description includes examples of one or more embodiments. Of course, it is not possible to describe all possible combinations of components or methods for the purpose of describing the above embodiments, but those of ordinary skill in the art should recognize that the various embodiments can be further combined and arranged. Therefore, the embodiments described herein are intended to cover all such changes, modifications, and variations that fall within the scope of the appended claims. In addition, with respect to the term "comprising" used in the specification or claims, this term is inclusive in a manner similar to the term "including", as is explained when "including" is used as a transitional word in a claim. In addition, any use of the term "or" in the specification or claims is intended to mean "non-exclusive or".

[0215] Those skilled in the art can also understand that the various illustrative logical blocks, units, and steps listed in the embodiments of the present invention can be implemented by electronic hardware, computer software, or a combination of both. To clearly show the interchangeability of hardware and software, the above various illustrative components, units, and steps have generally described their functions. Whether such functions are implemented by hardware or software depends on the specific application and the design requirements of the entire system. Those skilled in the art can use various methods to implement the described functions for each specific application, but such implementation should not be construed as exceeding the scope of protection of the embodiments of the present invention.

[0216] The various illustrative logical blocks or units described in the embodiments of the present invention can be implemented or operated to perform the described functions by a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination of the above designs. The general-purpose processor can be a microprocessor, and optionally, the general-purpose processor can also be any conventional processor, controller, microcontroller, or state machine. The processor can also be implemented by a combination of computing devices, such as a digital signal processor and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a digital signal processor core, or any other similar configuration.

[0217] In the embodiments of the present invention, the steps of the methods or algorithms described can be directly embedded in hardware, software modules executed by a processor, or a combination of the two. The software modules can be stored in a RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from the storage medium and write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and the storage medium can be provided in an ASIC, and the ASIC can be provided in a user terminal. Optionally, the processor and the storage medium can also be provided in different components of the user terminal.

[0218] In one or more exemplary designs, the above-described functions in the embodiments of the present invention can be implemented in hardware, software, firmware, or any combination of the three. If implemented in software, these functions can be stored on a computer-readable medium or transmitted on a computer-readable medium in the form of one or more instructions or codes. The computer-readable medium includes a computer storage medium and a communication medium that facilitates the transfer of a computer program from one place to another. The storage medium can be any available medium accessible by a general or special computer. For example, such a computer-readable medium can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM, or other optical disk storage, magnetic disk storage, or other magnetic storage devices, or any other medium that can be used to carry or store program code in the form of instructions or data structures and other forms readable by a general or special computer, or a general or special processor. In addition, any connection can be appropriately defined as a computer-readable medium. For example, if software is transmitted from a website, server, or other remote resource via a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless means such as infrared, wireless, and microwave, it is also included in the defined computer-readable medium. The disks (disk) and discs (disc) include compact discs, laser discs, optical discs, DVDs, floppy disks, and Blu-ray discs. Disks usually reproduce data magnetically, while discs usually reproduce data optically by laser. The above combinations can also be included in the computer-readable medium.

[0219] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. The autonomous navigation method for the Mars landing phase integrating point and line features is characterized in that: The method comprises the following steps: Obtain measurements of the Martian surface; Analyzing point features and line features in the measurement information to construct a measurement model; constructing a state model based on the measurement information and the dynamic model; Establishing an autonomous navigation model using the measurement model and the state model; The navigation state is updated through extended Kalman filtering according to the autonomous navigation model.

2. The Mars landing segment autonomous navigation method integrating point and line features according to claim 1, characterized in that: The step of analyzing the point features and line features in the measurement information to construct a measurement model comprises: Perform point feature pose estimation; Perform line feature pose estimation; The measurement model is constructed through point feature pose estimation and line feature pose estimation.

3. The Mars landing segment autonomous navigation method integrating point and line features according to claim 1, characterized in that: The measurement model is specifically: in, For the camera at t k Moment, for landmarks j The measurement of; f is the focal length of the camera; is the noise-free measurement; is the measurement noise, and its covariance matrix is is the vector from the spaceborne camera to the landmark.

4. The Mars landing segment autonomous navigation method integrating point and line features according to claim 1, characterized in that: The autonomous navigation model is specifically: Among them, r k is a vector block that integrates point and line measurement information; z imu =(a imu ,ω imu ) is the inertial measurement unit of IMU, t k is the k moment, in seconds; is the error state vector at time k.

5. The Mars landing segment autonomous navigation method integrating point and line features according to claim 1, characterized in that: The updating of the navigation state by using an extended Kalman filter according to the autonomous navigation model comprises: The measurement model is linearized. Linearizing the state model; Calculate the propagation of navigation state error covariance; By extending the Kalman filter algorithm, combining the linearized measurement model and the linearized state model, updating the navigation state and state covariance, the optimal estimation of the spacecraft state is achieved, thereby completing autonomous navigation.

6. The autonomous navigation device for the Mars landing stage integrating point and line features is characterized in that: include: A measurement unit, used to obtain measurement information of the Martian surface; An analysis unit, used for analyzing point features and line features in the measurement information to construct a measurement model; A construction unit, used for constructing a state model according to the measurement information and the dynamic model; A fusion unit, used to establish an autonomous navigation model using the measurement model and the state model; The navigation unit is used to update the navigation state through the extended Kalman filter according to the autonomous navigation model.

7. The Mars landing segment autonomous navigation device integrating point and line features according to claim 6, characterized in that: The analysis unit comprises: Perform point feature pose estimation; Perform line feature pose estimation; The measurement model is constructed through point feature pose estimation and line feature pose estimation.

8. The Mars landing segment autonomous navigation device integrating point and line features according to claim 6, characterized in that: The measurement model is specifically: in, For the camera at t k Moment, for landmarks j The measurement of; f is the focal length of the camera; is the noise-free measurement; is the measurement noise, and its covariance matrix is is the vector from the spaceborne camera to the landmark.

9. The Mars landing segment autonomous navigation device integrating point and line features according to claim 6, characterized in that: The autonomous navigation model is specifically: Among them, r k is a vector block that integrates point and line measurement information; z imu =(a imu ,ω imu ) is the inertial measurement unit of IMU, t k is the k moment, in seconds; is the error state vector at time k.

10. The Mars landing segment autonomous navigation device integrating point and line features according to claim 6, characterized in that: The navigation unit comprises: The measurement model is linearized. Linearizing the state model; Calculate the propagation of navigation state error covariance; By extending the Kalman filter algorithm, combining the linearized measurement model and the linearized state model, updating the navigation state and state covariance, the optimal estimation of the spacecraft state is achieved, thereby completing autonomous navigation.

Images