Self-correction state forecasting and fusion estimation method for Mars pneumatic auxiliary orbit descending model

Abstract

The invention discloses a self-correction state forecasting and fusion estimation method for a Mars pneumatic auxiliary orbit descending model, and belongs to the technical field of deep space exploration. The implementation method comprises the following steps: integrating various perturbation of a detector, and establishing a dynamic model suitable for long-term orbit extrapolation; the orbiter observes a natural large landmark on the surface of the Mars by using a navigation camera to obtain azimuth information of the landmark; when the detector is far away from the Mars, azimuth information of the Mars is obtained through the navigation camera; when the distance is closer to the Mars, observing a satellite surface landmark, and acquiring azimuth information of the landmark; the detector and the orbiter measure the relative distance and orientation through radio, and the measurement information obtained through fusion corrects the extrapolation result of the orbit. When the detector is located in the atmosphere, an accelerometer is used for measuring the external force acceleration of the detector, and inertial navigation is achieved. The atmospheric density is corrected by using the dynamic pressure and the heat flux density, and re-recursion is performed on the state of the detector, so that the state estimation precision of the detector in the Mars atmosphere is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to a Mars aerodynamic assisted deorbit model self-correction state prediction and fusion estimation method, which is particularly suitable for autonomous navigation tasks of a deep space probe during a Mars atmosphere assisted deorbit process, and belongs to the technical field of deep space exploration. BACKGROUND

[0002] In a Mars precise landing and sample return mission, fuel is a key factor affecting mission design. In order to save fuel for more scientific exploration, researchers have proposed a method of using Mars atmosphere for assisted deorbit. This method uses the resistance of the Mars atmosphere to change the speed of the probe, so that the probe reaches the predetermined orbit, which can significantly reduce fuel demand, reduce the launch cost of the probe, and increase the effective payload of the probe. During the atmospheric assisted orbit transfer, the probe passes through the high-altitude thin atmosphere several times, thereby gradually reducing the apogee height. The mission lasts for several months to a year, and an accurate dynamic model is the guarantee for the successful implementation of the atmospheric assisted orbit transfer mission. At the same time, the atmospheric density on the surface of Mars is affected by complex environmental factors and has strong uncertainty. Existing atmospheric density data and models cannot accurately calculate the atmospheric density at the location of the probe, which brings difficulties to the implementation of the atmospheric assisted orbit transfer mission. In the Mars atmospheric assisted orbit transfer mission, high-precision state estimation is a prerequisite for mission success. In order to achieve high-precision atmospheric assisted orbit transfer mission, an accurate atmospheric assisted deorbit state prediction and estimation method is needed to ensure the accuracy of the deorbit mission.

[0003] So far, Odyssey and MRO, which have successfully performed atmospheric assisted deorbit, have adopted an autonomous navigation scheme combining dead reckoning and radio ranging, which has realized the circumnavigation of Mars. However, dead reckoning cannot correct the initial error, combined with the random drift and error of the inertial measurement unit, external environmental disturbances and other factors, leading to the divergence of state estimation error over time. The introduction of radio ranging can correct the error of dead reckoning and improve the accuracy of state estimation. However, when the communication between the Earth and the probe is blocked by Mars, only dead reckoning can be relied on for navigation, which cannot meet the demand for autonomous navigation accuracy in the atmospheric assisted deorbit mission.

[0004] The Curiosity rover carried a Mars Entry Data Analysis System (MEDA) for the first time, which can measure the stagnation pressure through multiple pressure sensors. Since the measured dynamic pressure contains the relative velocity of the probe and the information of the Mars atmospheric density, the measured dynamic pressure can be used as a navigation observation when the probe is in the Mars atmosphere during the Mars atmospheric assisted descent, and an autonomous navigation scheme can be constructed by combining the inertial navigation. However, the high uncertainty of the atmospheric density will cause the inertial navigation error to increase rapidly, and affect the navigation accuracy. When the probe is outside the atmosphere and the radio communication fails, the lack of correction of the dynamic pressure measurement, the probe still uses the dead reckoning, and the state error will still diverge. Therefore, the high-reliability and high-precision Mars atmospheric assisted descent state prediction method and autonomous navigation scheme still need further research. SUMMARY

[0005] In view of the problem that the long-time orbit extrapolation accuracy is low during the Mars atmospheric assisted descent, the high uncertainty of the atmospheric density will cause the state prediction accuracy in the atmosphere to be low, and the state estimation error will diverge, the main purpose of the present application is to provide a Mars aerodynamic assisted descent model self-correcting state prediction and fusion estimation method. On the basis of comprehensively considering the multiple perturbations received by the probe, the real-time is ensured, and the orbit extrapolation accuracy is improved as much as possible. The measurement information obtained is used to correct the orbit extrapolation result, and the state estimation accuracy is improved. When the probe is in the atmosphere, the accelerometer is used to measure the external acceleration of the probe, and the inertial navigation is realized. The Mars Entry Data Analysis System (MEDA) and the temperature measuring probe are used to measure the stagnation dynamic pressure and the heat flux density of the probe, the atmospheric density is corrected by using the dynamic pressure and the heat flux density, and the state of the probe is re-repeated, and the state estimation accuracy of the probe in the Mars atmosphere is improved.

[0006] The purpose of the present application is realized by the following technical solutions.

[0007] The application discloses a Mars aerodynamic assisted deorbit model self-correction state prediction and fusion estimation method, which establishes a dynamic model suitable for long-term orbit extrapolation on the basis of comprehensively considering various perturbations suffered by a probe, ensures real-time, and improves the precision of orbit extrapolation as much as possible. The orbiter uses a navigation camera to observe natural large land marks on the surface of Mars to obtain the azimuth information of the land marks; when the probe is far away from Mars, the navigation camera of the probe observes the optical center of Mars to obtain the azimuth information of Mars; when the probe is close to Mars, the navigation camera of the probe observes the star table land mark to obtain the azimuth information of the land mark; meanwhile, the probe and the orbiter measure the relative distance and the azimuth through radio, and correct the results of orbit extrapolation by using the obtained measurement information to improve the precision of state estimation. When the probe is in the atmosphere, an accelerometer is used to measure the external acceleration of the probe to realize inertial navigation. The uncertainty of atmospheric density causes a large error of inertial navigation, in order to overcome the influence of the uncertainty of atmospheric density on inertial navigation, a Mars entry atmosphere data measurement system and a temperature measurement probe are used to measure the stagnation pressure and the heat flux density of the probe, the atmospheric density is corrected by using the stagnation pressure and the heat flux density, and the state of the probe is re-recalculated to improve the state estimation precision of the probe in the Mars atmosphere.

[0008] The Mars aerodynamic assisted deorbit model self-correction state prediction and fusion estimation method disclosed by the application comprises the following steps:

[0009] Step one, a high-precision probe dynamic model of atmospheric assisted variable orbit is established by comprehensively considering the Mars non-spherical gravity perturbation, three-body gravity perturbation, solar pressure and aerodynamic force, so as to facilitate orbit extrapolation in step three and improve the precision of long-time orbit extrapolation.

[0010] The Mars non-spherical perturbation potential function is expanded into the following spherical harmonic function form in the Mars fixed coordinate system

[0011]

[0012] And is a normalized Mars gravitational potential coefficient, which is obtained according to the GMM model; l and m are the order of the Mars gravitational potential coefficient; μ M is a Mars gravitational constant; R M is a Mars equatorial radius; is a normalized Legendre polynomial; is a normalized associated Legendre polynomial; is the length of the probe position vector r, and λ are the latitude and longitude of the probe in the Mars fixed coordinate system respectively, and satisfy

[0013]

[0014] The non-spherical perturbation acceleration of Mars is

[0015]

[0016] in, Let C be the gradient of the Martian nonspherical perturbation potential function with respect to the rectangular coordinates of the Martian fixed coordinate system. In actual calculations, it is necessary to transform the perturbation acceleration of the Martian fixed coordinate system to the inertial coordinate system. MCF_MCI It is the transformation matrix from the Mars J2000.0 Earth equatorial coordinate system to the Mars fixed coordinate system.

[0017] In the Mars J2000.0 Earth equatorial coordinate system, the perturbation acceleration caused by the Sun's gravity on the probe is:

[0018]

[0019] Where, μ s Represents the gravitational constant of the Sun; r and r s These represent the position vectors of the probe and the sun, respectively.

[0020] In the Mars J2000.0 Earth equatorial coordinate system, the perturbation acceleration caused by solar radiation pressure on the probe is:

[0021]

[0022] Where r and r s S and n represent the position vectors of the probe and the sun, respectively; S / n represents the probe's surface-to-mass ratio; C R Represents the solar radiation coefficient; c represents the speed of light; L s represents solar luminosity; K represents the solar visibility coefficient at the location of the detector.

[0023] The position vector r of the Sun in the Mars J2000.0 Earth equatorial coordinate system s The calculation is performed using the following method. In the heliocentric J2000.0 ecliptic system, the average orbital elements of Mars are...

[0024]

[0025] Where D is one astronomical unit; T is the Julian century number.

[0026] The near angle ω and the true near angle f are calculated by the following formula.

[0027]

[0028] Where M satisfies the following equation

[0029] Thus, the six orbital elements of Mars in the heliocentric J2000.0 ecliptic inertial frame are obtained. Based on the method of converting orbital elements to state quantities in Cartesian coordinates, the position vector r of Mars in the heliocentric J2000.0 ecliptic inertial frame can be obtained. Mars_Ecl Mars' position vector r in the heliocentric J2000.0 Earth equatorial coordinate system. Mars_Eq for

[0030]

[0031] C Ecl_Eq This is the transformation matrix from the heliocentric J2000.0 Earth equatorial coordinate system to the heliocentric J2000.0 ecliptic inertial frame. The position vector r of the Sun in the Mars J2000.0 Earth equatorial coordinate system. s for

[0032] r s =-r Mars_Eq (11)

[0033] In the Mars J2000.0 Earth equatorial coordinate system, what is the drag acceleration 'a' experienced by the probe? D and lift acceleration a L Calculated by the following formula

[0034]

[0035] Where n represents the mass of the probe, D and L are the atmospheric drag and lift forces acting on the probe in the fixed Martian coordinate system, respectively, and C... MCF_MCI This is the coordinate transformation matrix from the Earth equatorial coordinate system (Mars J2000.0) to the Mars fixed coordinate system.

[0036] The Martian atmosphere rotates with Mars, meaning it is stationary relative to Mars. In a fixed Martian coordinate system, the probe's position r... MCF and velocity vector v MCF for

[0037]

[0038] Where r and v are the position vector and velocity vector of the probe in the Mars J2000.0 Earth equatorial coordinate system, respectively; ω M Let be the Martian rotation angular velocity. In the fixed Martian coordinate system, the drag vector D and lift vector L experienced by the probe can be calculated by the following formula.

[0039]

[0040] Among them, C D and C L These represent the probe's drag coefficient and lift coefficient, respectively; S is the probe's reference area; and ρ is the Martian atmospheric density.

[0041] The Martian atmospheric density ρ is calculated using an exponential model, i.e.

[0042]

[0043] Where ρ0 is the reference atmospheric density, h0 is the reference altitude, and h s The altitude is the proportional height; h is the probe's current altitude. Mars is an ellipsoid, and h is calculated using the following formula.

[0044]

[0045] Among them, R M e is the radius of the Martian equator; ECCM r is the oblateness of Mars. MCF =||r MCF || represents the distance from the detector's center of fire; λ represents the latitude, satisfying...

[0046]

[0047] In the Mars J2000.0 Earth equatorial coordinate system, the probe's dynamic vector equations are expressed in the following form.

[0048]

[0049] Where r is the position vector of the detector, a M For the non-spherical gravitational perturbation experienced by the detector, a T For the solar gravitational perturbation experienced by the probe, a R a is the solar radiation pressure experienced by the detector. D and a L h represents the drag acceleration and lift acceleration experienced by the probe, respectively. air R represents the altitude of the Martian atmosphere. M This is the radius of the Martian equator.

[0050] Step 2: Using the optical center of Mars and large landmarks on the Martian surface as camera observation targets, construct an observation model of the optical center of Mars and landmarks by comprehensively measuring noise; using the azimuth and distance between the orbiter and the probe as relative measurements, construct a relative measurement model by comprehensively measuring noise.

[0051] The probe and orbiter observe the line-of-sight vectors of Mars and landmarks using their cameras. These line-of-sight vectors are equivalent to azimuth and elevation angles. The relationship between the azimuth and elevation angles and the positions of the probe and orbiter is as follows:

[0052]

[0053] θ o This indicates the azimuth of the line of sight to the landmass observed by the orbiter. θ represents the elevation angle of the landmass observed by the orbiter. s This indicates the azimuth angle of the line of sight to the landmass observed by the probe. The x-axis represents the elevation angle of the landmass observed by the probe. o ,y o ,z o ] T Indicates the relative position of the landmass observed by the orbiter to the orbiter, [x s ,y s ,z s ] T This indicates the relative position of the Martian photocenter or landmark observed by the probe to the probe.

[0054] Considering measurement noise, a camera measurement model is established as follows:

[0055]

[0056] Where ε os To measure noise, the standard deviation is σ. os The noise is Gaussian white noise, and x is a state variable.

[0057] The orbiter and the probe measure their relative distance, azimuth, and elevation angles via radio, and their positional relationship with the probe and orbiter is as follows:

[0058]

[0059] θ r This indicates the line-of-sight azimuth angle between the probe and the orbiter. d represents the pitch angle between the probe and the orbiter. r [x] represents the relative distance between the probe and the orbiter. r ,y r ,z r ] T This indicates the relative position between the probe and the orbiter.

[0060] Considering measurement noise, a relative measurement model is established as follows:

[0061]

[0062] Where ε r The relative measurement noise is σ. r Gaussian white noise.

[0063] Step 3: Integrate the camera measurement and relative measurement information from Step 2 to construct a Mars atmospheric-assisted descent autonomous navigation observation model, ensuring the reliability of the probe's autonomous navigation and improving the accuracy of Mars atmospheric-assisted descent autonomous navigation.

[0064] Based on the sensor measurement model in equations (20) and (22) of step two, the Mars atmospheric-assisted descent navigation observation model is established as follows:

[0065]

[0066] When the probe is far from Mars, the target observed by the probe's camera is the photocenter of Mars; when the probe is close to Mars, the target observed by the probe's camera is the planetary landmarks.

[0067] Step 4: Based on the detector dynamics model in Step 1 and the navigation observation model in Step 3, the navigation state variables include the position and velocity of the detector and the position and velocity of the orbiter. The camera observation information and relative measurement information are fused together, and the extended Kalman filter (EKF) is used to estimate the state variables.

[0068] Based on the detector dynamics model established by equation (18) and the navigation observation model established by equation (23), the navigation state variables include the position and velocity of the detector and the position and velocity of the orbiter. The extended Kalman filter (EKF) is used to estimate the state variables.

[0069] One-step prediction of state variables and state error covariance matrix based on the dynamic model of equation (18)

[0070]

[0071] Where f represents the dynamic model shown in equation (18), This represents the state estimate at time k-1. P represents the state prediction value at time k. k-1 Let Q be the state error covariance matrix at time k-1. k-1 For the process noise at time k-1, P k|k-1 Let Φ be the predicted value of the state error covariance matrix at time k. k-1 Let be the state transition matrix at time k-1.

[0072]

[0073] Let I be the identity matrix and F be the dynamic Jacobian matrix, then we have Δt is the navigation observation time interval.

[0074] Based on the autonomous navigation observation model of equation (23), the observation Jacobian matrix H and residual z corresponding to the navigation observation scheme are given.

[0075] The subsequent state update is performed using the observed Jacobian matrix H and the residual z, and the update equation is as follows:

[0076]

[0077] Where R k The variance matrix of the navigation observation error at time k is determined by the measurement noise of each sensor in step two; K k Let I be the filter gain at time k, and I be the identity matrix. Let P be the state estimate at time k. k Let be the state estimation error covariance matrix at time k.

[0078] Step 5: Integrate the stagnation point dynamic pressure and heat flux density measurement information to perform online self-calibration of the Martian atmospheric density and re-evaluate the probe's state to improve the accuracy of the probe's state prediction within the atmosphere and provide a guarantee for high-precision state estimation.

[0079] When the detector is located within the atmosphere, the stagnation point dynamic pressure and heat flux density are measured. The detector's stagnation point dynamic pressure and heat flux density are:

[0080]

[0081] q represents the stagnation point dynamic pressure value of the detector. The detector's heat flux density is represented by ρ, where ρ is the atmospheric density and v is the atmospheric density. rel k is the velocity of the probe relative to the surface of Mars. q R represents the heat flux coefficient. n This indicates the curvature of the probe's head.

[0082] Considering measurement noise, a measurement model for dynamic pressure and heat flux density is established as follows:

[0083]

[0084] Where ε a To measure noise, the standard deviation is σ. a Gaussian white noise.

[0085] The predicted distance from the detector's center of fire Substituting into equation (16), we can obtain the altitude prediction value h. pre , and then h pre Substituting into the density model of equation (15), the predicted density value ρ is obtained. pre .in, This is the detector's position prediction value, and it is the state prediction value. The first three components. The observation Jacobian matrix H required for atmospheric density estimation. ρ With density recurrence matrix F ρ Calculation as follows

[0086]

[0087] Similarly, the atmospheric density at the current moment can be estimated using equations (24), (25), and (27), thus correcting the atmospheric density. The corrected atmospheric density is ρ. est Substitute it into equation (14) to modify the dynamic model, and then use the modified dynamic model to re-predict the state in equation (24) to complete the subsequent state estimation process.

[0088] Thus, the self-correction state prediction and fusion estimation method of the Mars aerodynamic-assisted descent model has been completed, ensuring the accuracy of probe state prediction under the condition of high uncertainty in atmospheric density, improving the accuracy of probe state estimation during the Mars atmospheric-assisted descent process, and realizing high-precision autonomous navigation for Mars atmospheric-assisted descent.

[0089] Beneficial effects:

[0090] 1. The self-correction state prediction and fusion estimation method for the Mars aerodynamic descent model disclosed in this invention establishes a high-precision orbit recursion model, considering multiple perturbations such as solar radiation pressure, solar gravity, and non-spherical gravity of Mars, to ensure the accuracy of long-term orbit recursion during the Mars aerodynamic descent process, thus providing a guarantee for the accuracy of autonomous navigation.

[0091] 2. The Mars aerodynamic-assisted descent model self-correction state prediction and fusion estimation method disclosed in this invention adds dynamic pressure and heat flux density measurement information in the atmosphere, avoiding the problem of poor state prediction accuracy caused by the high uncertainty of atmospheric density in the atmosphere, and effectively ensuring navigation accuracy in the atmosphere.

[0092] 3. The Mars aerodynamic-assisted descent model self-correction state prediction and fusion estimation method disclosed in this invention utilizes relative measurement information of the orbiter and observation information of Mars and its surface landmarks for autonomous navigation. It makes full use of available resources and information and can still achieve high-precision autonomous navigation even when the orbiter's state is unknown.

[0093] 4. In order to overcome the influence of atmospheric density uncertainty on inertial navigation, the self-correction state prediction and fusion estimation method of Mars aerodynamic-assisted descent model disclosed in this invention uses a Mars atmospheric entry data measurement system and a temperature probe to measure the dynamic pressure and heat flux density of the probe's stagnation point. The dynamic pressure and heat flux density are used to correct the atmospheric density, and the probe's state is re-estimated to improve the accuracy of the probe's state estimation in the Martian atmosphere. Attached Figure Description

[0094] Figure 1 This is a flowchart illustrating the self-correction state prediction and fusion estimation method for the Mars aerodynamic-assisted orbit descent model of the present invention.

[0095] Figure 2 It refers to the three-dimensional orbit of the detector and the orbiter in the specific implementation.

[0096] Figure 3 This refers to the detector state estimation error in the specific implementation, where (a) is the estimation error of position x, y, z; and (b) is the estimation error of velocity v. x ,v y ,v z The estimation error. Detailed Implementation

[0097] To better illustrate the purpose and advantages of the present invention, the invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0098] To verify the feasibility of the method, the initial states of the probe and orbiter in this example are shown in Table 1. The Mars orbiter with an orbital altitude of 350 km was selected as the target for relative measurement.

[0099] like Figure 1 As shown in the example, the self-correcting state prediction and fusion estimation method for the Mars aerodynamically assisted orbit descent model disclosed in this example has the following specific implementation steps:

[0100] Step 1: Integrate the non-spherical gravitational perturbations of Mars, the three-body gravitational perturbations, solar radiation pressure, and aerodynamic forces to establish a dynamic model for a high-precision atmospheric-assisted orbital maneuvering probe. This will facilitate orbit extrapolation in the subsequent Step 3 and improve the accuracy of long-term orbital extrapolation.

[0101] The non-spherical perturbation potential function of Mars can be expanded into the following spherical harmonic function form in the fixed coordinate system of Mars.

[0102] and The normalized Martian gravitational potential coefficients are obtained from the 80×80 order GMM (Goddard Mars Model), where l and m represent the order of the Martian gravitational potential coefficients; and R is the Martian equatorial radius. M =3397km; μ M The gravitational constant of Mars; For normalized Legendre polynomials; It is a normalized associative Legendre polynomial; Let r be the length of the position vector. λ and λ represent the latitude and longitude of the probe in the fixed coordinate system on Mars, respectively, satisfying the following conditions:

[0103]

[0104] The non-spherical perturbation acceleration of Mars is

[0105]

[0106] in, Let C be the gradient of the Martian nonspherical perturbation potential function with respect to the Cartesian coordinates of the Martian fixed coordinate system. In actual calculations, it is necessary to transform the perturbation acceleration of the Martian fixed coordinate system to the inertial coordinate system. MCF_MCI It is the transformation matrix from the Mars J2000.0 Earth equatorial coordinate system to the Mars fixed coordinate system.

[0107] In the Mars J2000.0 Earth equatorial coordinate system, the perturbation acceleration caused by the Sun's gravity on the probe is:

[0108]

[0109] Where, μ s Represents the solar gravitational constant; r and r s These represent the position vectors of the detector.

[0110] In the Mars J2000.0 Earth equatorial coordinate system, the perturbation acceleration caused by solar radiation pressure on the probe is:

[0111]

[0112] Where r and r s S and n represent the position vectors of the probe and the sun, respectively; S / n represents the probe's surface-to-mass ratio; C R =1 is the solar radiation coefficient; c is the speed of light; L s =3.823×10 26 K represents the solar luminosity; K=1 represents the solar visibility coefficient at the location of the detector.

[0113] In the Mars J2000.0 Earth equatorial coordinate system, what is the drag acceleration 'a' experienced by the probe? D and lift acceleration a L It can be calculated by the following formula

[0114]

[0115] Where n represents the mass of the probe; D and L represent the atmospheric drag and lift experienced by the probe in the fixed Martian coordinate system, respectively; and C represents the mass of the probe. MCF_MCI This is the coordinate transformation matrix from the Earth equatorial coordinate system (Mars J2000.0) to the Mars fixed coordinate system.

[0116] In summary, in the Mars J2000.0 Earth equatorial inertial frame, the probe's dynamic vector equations can be expressed as follows:

[0117]

[0118] Where, μ M Let r be the Martian gravitational constant, r be the probe's position vector, and a be the position vector. M For the non-spherical gravitational perturbation experienced by the detector, aT For the solar gravitational perturbation experienced by the probe, a R a is the solar radiation pressure experienced by the detector. D and a L h represents the drag acceleration and lift acceleration experienced by the probe, respectively. air R represents the altitude of the Martian atmosphere. M This is the radius of the Martian equator.

[0119] In equations (4) and (5), the solar position vector r in the Mars J2000.0 Earth equatorial coordinate system is... s The following method was used to obtain the average orbital elements of Mars in the heliocentric J2000.0 ecliptic system:

[0120]

[0121] Where, D = 1.495978707 × 10 11 m is an astronomical unit; T is the Julian century number.

[0122] The near angle ω and the true near angle f can be calculated by the following formula.

[0123]

[0124] Where M satisfies the following equation

[0125] Thus, we can obtain the six orbital elements of Mars in the heliocentric J2000.0 ecliptic inertial frame. Using the method for converting orbital elements to state variables in Cartesian coordinates, we can obtain the position vector r of Mars in the heliocentric J2000.0 ecliptic inertial frame. Mars_Ecl Mars' position vector r in the heliocentric J2000.0 Earth equatorial coordinate system. Mars_Eq for

[0126]

[0127] C Ecl_Eq This is the transformation matrix from the heliocentric J2000.0 Earth equatorial inertial frame to the heliocentric J2000.0 ecliptic inertial frame. Therefore, the position vector r of the Sun in the Mars J2000.0 Earth equatorial coordinate system... s for

[0128] r s =-r Mars_Eq (13)

[0129] D and L in equation (6) are obtained by the following method

[0130]

[0131] Among them, C Dand C L These represent the probe's drag coefficient and lift coefficient, respectively; S is the probe's reference area; ρ is the Martian atmospheric density; r MCF v is the position vector of the detector. MCF is the velocity vector of the detector.

[0132] The position vector and velocity vector of the detector can be calculated by the following formula.

[0133]

[0134] Where r and v are the position vector and velocity vector of the probe in the Mars J2000.0 Earth equatorial coordinate system, respectively; ω M This is the angular velocity of Mars' rotation.

[0135] The density of the Martian atmosphere was calculated using an exponential model, i.e.

[0136]

[0137] Where ρ0 is the reference atmospheric density, h0 is the reference altitude, and h s ρ is the proportional altitude; h is the probe's current altitude. Mars is an ellipsoid, and h is calculated using the following formula.

[0138]

[0139] Among them, R M =3397km is the radius of the Martian equator; e ECCM =0.005886007555525 is the oblateness of Mars; r MCF =||r MCF || represents the distance from the detector's center of fire; λ represents the latitude, satisfying...

[0140]

[0141] Step 2: Using the optical center of Mars and large landmarks on the Martian surface as camera observation targets, construct an observation model of the optical center of Mars and landmarks by comprehensively measuring noise; using the azimuth and distance between the orbiter and the probe as relative measurements, construct a relative measurement model by comprehensively measuring noise.

[0142] The probe and orbiter observe the line-of-sight vectors of Mars and landmarks using their cameras. These line-of-sight vectors are equivalent to azimuth and elevation angles. The relationship between the azimuth and elevation angles and the positions of the probe and orbiter is as follows:

[0143]

[0144] θ o This indicates the azimuth of the line of sight to the landmass observed by the orbiter. θ represents the elevation angle of the landmass observed by the orbiter.s This indicates the azimuth angle of the line of sight to the landmass observed by the probe. The x-axis represents the elevation angle of the landmass observed by the probe. o ,y o ,z o ] T Indicates the relative position of the landmass observed by the orbiter to the orbiter, [x s ,y s ,z s ] T This indicates the relative position of the Martian photocenter or landmark observed by the probe to the probe.

[0145] Considering measurement noise, a camera measurement model is established as follows:

[0146]

[0147] Where ε os To measure noise, the standard deviation is σ. os The noise is Gaussian white noise, and x is a state variable.

[0148] The orbiter and the probe measure their relative distance, azimuth, and elevation angles via radio, and their positional relationship with the probe and orbiter is as follows:

[0149]

[0150] θ r This indicates the line-of-sight azimuth angle between the probe and the orbiter. d represents the pitch angle between the probe and the orbiter. r [x] represents the relative distance between the probe and the orbiter. r ,y r ,z r ] T This indicates the relative position between the probe and the orbiter.

[0151] Considering measurement noise, a relative measurement model is established as follows:

[0152]

[0153] Where ε r The relative measurement noise is σ. r Gaussian white noise.

[0154] Step 3: Integrate multi-source measurement information from camera measurements and relative measurements to construct an observation model for Mars atmospheric-assisted descent autonomous navigation, ensuring the reliability of the probe's autonomous navigation and improving the accuracy of Mars atmospheric-assisted descent autonomous navigation.

[0155] Based on the sensor measurement model in equations (20) and (22) of step two, the Mars atmospheric-assisted descent navigation observation model is established as follows:

[0156]

[0157] When the probe is far from Mars, the target observed by the probe's camera is the photocenter of Mars; when the probe is close to Mars, the target observed by the probe's camera is the planetary landmarks.

[0158] Step 4: Based on the detector dynamics model in Step 1 and the navigation observation model in Step 3, the navigation state variables include the position and velocity of the detector and the position and velocity of the orbiter. The camera observation information and relative measurement information are fused, and the extended Kalman filter (EKF) is used to estimate the state variables.

[0159] One-step prediction of state variables and state error covariance matrix based on the dynamic model of equation (7)

[0160]

[0161] Where f represents the dynamic model shown in equation (7), This represents the state estimate at time k-1. P represents the state prediction value at time k. k-1 Let Q be the state error covariance matrix at time k-1. k-1 For the process noise at time k-1, P k|k-1 Let Φ be the predicted value of the state error covariance matrix at time k. k-1 Let be the state transition matrix at time k-1.

[0162]

[0163] Let I be the identity matrix and F be the dynamic Jacobian matrix, then we have Δt is the navigation observation time interval.

[0164] Based on the autonomous navigation observation model of equation (23), the observation Jacobian matrix H and residual z corresponding to the navigation observation scheme are given.

[0165] The subsequent state update is performed using the observed Jacobian matrix H and the residual z, and the update equation is as follows:

[0166]

[0167] Where R k The variance matrix of the navigation observation error at time k is determined by the measurement noise of each sensor in step two. k Let I be the filter gain at time k, and I be the identity matrix. Let P be the state estimate at time k. k Let be the state estimation error covariance matrix at time k.

[0168] Step 5: Integrate the stagnation point dynamic pressure and heat flux density measurement information to make online corrections to the Martian atmospheric density and re-evaluate the probe's state to improve the accuracy of the probe's state prediction within the atmosphere and provide a guarantee for high-precision state estimation.

[0169] When the detector is located within the atmosphere, the stagnation point dynamic pressure and heat flux density are measured. The detector's stagnation point dynamic pressure and heat flux density are:

[0170]

[0171] q represents the stagnation point dynamic pressure value of the detector. The detector's heat flux density is represented by ρ, where ρ is the atmospheric density and v is the atmospheric density. rel k is the velocity of the probe relative to the surface of Mars. q =1.9027×10 -4 R represents the heat flux coefficient. n =1 indicates the head curvature of the detector.

[0172] Considering measurement noise, a measurement model for dynamic pressure and heat flux density is established as follows:

[0173]

[0174] Where ε a To measure noise, the standard deviation is σ. a Gaussian white noise.

[0175] The predicted distance from the detector's center of fire Substituting into the height model of equation (17), the predicted height value h can be obtained. pre , will h pre Substituting into the density model of equation (16), the predicted density value ρ can be obtained. pre . This is the detector's position prediction value, and it is the state prediction value. The first three components. The observation Jacobian matrix H required for atmospheric density estimation. ρ With density recurrence matrix F ρ Calculation as follows

[0176]

[0177] Similarly, the atmospheric density at the current moment can be estimated according to equations (24), (25), and (27), achieving self-correction of atmospheric density. The corrected atmospheric density is ρ. est , will ρ estSubstituting into equation (14), the dynamic model is modified, and then the modified dynamic model is used to re-predict the state in equation (24) and complete the subsequent state estimation process.

[0178] Thus, the self-correction state prediction and fusion estimation method of the Mars aerodynamic-assisted descent model has been completed, ensuring the accuracy of probe state prediction under the condition of high uncertainty in atmospheric density, improving the accuracy of probe state estimation during the Mars atmospheric-assisted descent process, and realizing high-precision autonomous navigation for Mars atmospheric-assisted descent.

[0179] The parameter settings are shown in Table 1.

[0180] Table 1 Parameter Settings

[0181]

[0182] like Figure 3 As shown, using the Mars aerodynamic-assisted descent model self-correction state prediction and fusion estimation method disclosed in this invention, the probe's initial position is outside the atmosphere. Autonomous navigation is performed by fusing camera measurements and relative measurements. The probe's state estimation error converges rapidly, with a position estimation error of less than 1 km and a velocity estimation error of less than 0.5 m / s. This indicates that the dynamic recursive model established by this method has high accuracy and can meet the needs of long-term orbit recursion. Simultaneously, the navigation scheme fused with camera observations and relative measurements has good estimation accuracy. When the probe is inside the atmosphere, atmospheric density has high uncertainty. At this time, the state prediction accuracy is affected by atmospheric density deviation, resulting in a large error. Using the autonomous navigation method fused with relative measurements and camera observations, the state error initially diverges and then converges again, leading to unsatisfactory state estimation results. However, after online correction of atmospheric density, the problem of high atmospheric density uncertainty can be effectively addressed. State re-recursion prevents the divergence of state estimation errors caused by large state prediction errors upon atmospheric entry. The convergence effect of navigation errors is better within the atmosphere, resulting in higher reliability.

[0183] The above detailed description further illustrates the purpose, technical solution, and beneficial effects of the invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A self-correcting state prediction and fusion estimation method for a Mars aerodynamic-assisted orbit descent model, characterized by: Includes the following steps, Step 1: Integrate the non-spherical gravitational perturbations of Mars, the three-body gravitational perturbations, solar radiation pressure, and aerodynamic forces to establish a dynamic model for a high-precision atmospheric-assisted orbit-changing probe; this will facilitate orbit extrapolation in the subsequent Step 3 and improve the accuracy of long-term orbit extrapolation. Step 2: Using the optical center of Mars and large landmarks on the Martian surface as camera observation targets, construct an observation model of the optical center of Mars and landmarks by comprehensively measuring noise; using the azimuth and distance between the orbiter and the probe as relative measurements, construct a relative measurement model by comprehensively measuring noise. Step 3: Integrate the camera observation model and relative measurement model from Step 2 to construct a Mars atmosphere-assisted descent autonomous navigation observation model; Step 4: Based on the detector dynamics model in Step 1 and the navigation observation model in Step 3, the navigation state variables include the position and velocity of the detector and the position and velocity of the orbiter. The camera observation information and relative measurement information are fused, and the extended Kalman filter (EKF) is used to estimate the state variables.

2. The method as described in claim 1, characterized in that: It also includes step five, which integrates the stagnation point dynamic pressure and heat flux density measurement information to perform online self-calibration of the Martian atmospheric density and re-evaluate the probe's state to improve the accuracy of the probe's state prediction in the atmosphere and provide a guarantee for high-precision state estimation. Based on steps one through five, a self-correcting state prediction and fusion estimation method for the Mars aerodynamic-assisted descent model is completed. This ensures the accuracy of probe state prediction under conditions of high uncertainty in atmospheric density, improves the accuracy of probe state estimation during Mars atmospheric-assisted descent, and enables high-precision autonomous navigation during Mars atmospheric-assisted descent.

3. The method as described in claim 1, characterized in that: The process of establishing the dynamic model described in step one is as follows: The non-spherical perturbation potential function of Mars expands into the following spherical harmonic function form in the fixed coordinate system of Mars. and The normalized Martian gravitational potential coefficients are obtained from the GMM model; l,m represents the order corresponding to the Martian gravitational potential coefficients; μ M R is the gravitational constant of Mars. M This is the radius of the Martian equator; For normalized Legendre polynomials; It is a normalized associative Legendre polynomial; The length of the detector position vector r, λ and λ represent the latitude and longitude of the probe in the fixed coordinate system on Mars, respectively, satisfying the following conditions: The non-spherical perturbation acceleration of Mars is in, Let C be the gradient of the Martian nonspherical perturbation potential function with respect to the rectangular coordinates of the Martian fixed coordinate system. In actual calculations, it is necessary to transform the perturbation acceleration of the Martian fixed coordinate system to the inertial coordinate system. MCF_MCI It is the transformation matrix from the Mars J2000.0 Earth equatorial inertial frame to the Mars fixed coordinate system; In the Mars J2000.0 Earth equatorial inertial frame, the perturbation acceleration caused by the Sun's gravity on the probe is: Where, μ s Represents the gravitational constant of the Sun; r and r s These represent the position vectors of the probe and the sun, respectively. In the Mars J2000.0 Earth equatorial inertial frame, the perturbation acceleration caused by solar radiation pressure on the probe is: Where r and r s S and n represent the position vectors of the probe and the sun, respectively; S / n represents the probe's surface-to-mass ratio; C R Represents the solar radiation coefficient; c represents the speed of light; L s K represents the solar luminosity; K represents the solar visibility coefficient at the location of the detector. In the Mars J2000.0 Earth equatorial inertial frame, what is the drag acceleration 'a' experienced by the probe? D and lift acceleration a L It can be calculated by the following formula Where n represents the mass of the probe; D and L are the atmospheric drag vector and lift vector of the probe in the Mars fixed coordinate system, respectively; and C... MCF_MCI This is the coordinate transformation matrix from the Earth equatorial inertial frame J2000.0 to the Mars fixed coordinate system; In the Mars J2000.0 Earth equatorial inertial frame, the probe's dynamic vector equations are expressed as follows: Where, μ M Let r be the Martian gravitational constant, r be the probe's position vector, and a be the position vector. M For the non-spherical gravitational perturbation experienced by the detector, a T For the solar gravitational perturbation experienced by the probe, a R a is the solar radiation pressure experienced by the detector. D and a L h represents the drag acceleration and lift acceleration experienced by the probe, respectively. air R represents the altitude of the Martian atmosphere. M This is the radius of the Martian equator.

4. The method as described in claim 3, characterized in that: The solar position vector r s The following method was used to obtain the average orbital elements of Mars in the heliocentric J2000.0 ecliptic system: Where D is one astronomical unit; T is the Julian century number; The near angle ω and the true near angle f are calculated by the following formula. Where M satisfies the following equation Thus, the six orbital elements of Mars in the heliocentric J2000.0 ecliptic inertial frame are obtained. Based on the method for converting orbital elements to state quantities in Cartesian coordinates, the position vector r of Mars in the heliocentric J2000.0 ecliptic inertial frame can be obtained. Mars_Ecl Mars' position vector r in the Earth's equatorial inertial frame at heliocentric J2000.0 Mars_Eq for C Ecl_Eq This is the transformation matrix from the heliocentric J2000.0 Earth equatorial inertial frame to the heliocentric J2000.0 ecliptic inertial frame. Therefore, the position vector r of the Sun in the Mars J2000.0 Earth equatorial inertial frame is... s for r s ＝-r Mars_Eq (13) 5. The method as described in claim 3, characterized in that: In step one, equation (6), D and L are obtained through the following method. Among them, C D and C L These represent the probe's drag coefficient and lift coefficient, respectively; S is the probe's reference area; ρ is the Martian atmospheric density; and r is the probe's position vector. MCF and velocity vector v MCF for Where r and v are the position vector and velocity vector of the probe in the Mars J2000.0 Earth equatorial inertial frame, respectively; ω M This is the angular velocity of Mars' rotation; The Martian atmospheric density ρ is calculated using an exponential model, i.e. Where ρ0 is the reference atmospheric density, h0 is the reference altitude, and h s ρ is the proportional altitude; h is the probe's current altitude. Mars is an ellipsoid, and h is calculated using the following formula. Among them, R M e is the radius of the Martian equator; ECCM r is the oblateness of Mars. MCF =||r MCF || represents the distance from the detector's center of fire; λ represents the latitude, satisfying...

6. The method as described in claim 1, characterized in that: Step two describes a method for constructing an observation model of the Martian optical center and large Martian landmarks by using them as camera observation targets and comprehensively measuring noise. The probe and orbiter observe the line-of-sight vectors of Mars and landmarks using their cameras. These line-of-sight vectors are equivalent to azimuth and elevation angles. The relationship between the azimuth and elevation angles and the positions of the probe and orbiter is as follows: θ o This indicates the azimuth of the line of sight to the landmass observed by the orbiter. θ represents the elevation angle of the landmass observed by the orbiter. s This indicates the azimuth angle of the line of sight to the landmass observed by the probe. The x-axis represents the elevation angle of the landmass observed by the probe. o ,y o ,z o ] T Indicates the relative position of the landmass observed by the orbiter to the orbiter, [x s ,y s ,z s ] T This indicates the relative position of the Martian optical center or landmark observed by the probe to the probe itself; Considering measurement noise, a camera measurement model is established as follows: Where ε os To measure noise, the standard deviation is σ. os The noise is Gaussian white noise, and x is a state variable.

7. The method as described in claim 1, characterized in that: Step two describes a method for constructing a relative measurement model by using the azimuth and distance between the orbiter and the detector as relative measurements and integrating measurement noise. The orbiter and the probe measure their relative distance, azimuth, and elevation angles via radio, and their positional relationship with the probe and orbiter is as follows: θ r This indicates the line-of-sight azimuth angle between the probe and the orbiter. d represents the pitch angle between the probe and the orbiter. r [x] represents the relative distance between the probe and the orbiter. r ,y r ,z r ] T Indicates the relative position between the probe and the orbiter; Considering measurement noise, a relative measurement model is established as follows: Where ε r The relative measurement noise is σ. r Gaussian white noise.

8. The method as described in claim 1, characterized in that: The implementation method for step three is as follows: Based on equations (20) and (22) in step two, the following Martian atmospheric-assisted descent autonomous navigation observation model is established: When the probe is far from Mars, the target observed by the probe's camera is the photocenter of Mars; when the probe is close to Mars, the target observed by the probe's camera is the planetary landmarks.

9. The method as described in claim 1, characterized in that: The specific implementation method for step four is as follows: One-step prediction of state variables and state error covariance matrix based on the dynamic model of equation (7) Where f represents the dynamic model shown in equation (7), This represents the state estimate at time k-1. P represents the state prediction value at time k. k-1 Let Q be the state error covariance matrix at time k-1. k-1 For the process noise at time k-1, P k|k-1 Let Φ be the predicted value of the state error covariance matrix at time k. k-1 Let be the state transition matrix at time k-1; where Let I be the identity matrix and F be the dynamic Jacobian matrix, then we have Δt is the navigation observation time interval; Based on the autonomous navigation observation model of equation (23), the observation Jacobian matrix H and residual z corresponding to the navigation observation scheme are given; The subsequent state update is performed using the observed Jacobian matrix H and the residual z, and the update equation is as follows: Where R k The variance matrix of the navigation observation error at time k is determined by the measurement noise of each sensor in step two; K k Let I be the filter gain at time k, and I be the identity matrix. Let P be the state estimate at time k. k Let be the state estimation error covariance matrix at time k.

10. The method as described in claim 2, characterized in that: The implementation method for step five is as follows: When the detector is located within the atmosphere, the stagnation point dynamic pressure and heat flux density are measured. The detector's stagnation point dynamic pressure and heat flux density are: q represents the stagnation point dynamic pressure value of the detector. The detector's heat flux density is represented by ρ, where ρ is the atmospheric density and v is the atmospheric density. rel k is the velocity of the probe relative to the surface of Mars. q R represents the heat flux coefficient. n Indicates the curvature of the probe's head; Considering measurement noise, a measurement model for dynamic pressure and heat flux density is established as follows: Where ε a To measure noise, the standard deviation is σ. a Gaussian white noise; The predicted distance from the detector's center of fire Substituting into equation (17), we obtain the altitude prediction value h. pre , and then h pre Substituting into the density model of equation (16), the predicted density value ρ can be obtained. pre . This is the detector's position prediction value, and it is the state prediction value. The first three components; the observation Jacobian matrix H required for atmospheric density estimation. ρ With density recurrence matrix F ρ Calculation as follows Similarly, the atmospheric density at the current moment is estimated according to equations (24), (25), and (27) to correct the atmospheric density; the corrected atmospheric density is ρ. est , will ρ est Substitute into equation (14) to modify the dynamic model, and then use the modified dynamic model to re-predict the state of equation (24) and complete the subsequent state estimation process.

Images