GNSS-S radar beam resource optimization method for on-orbit batch ship tracking

Abstract

The invention relates to a GNSS-S radar beam resource optimization method for on-orbit batch ship tracking. The method comprises the following steps: acquiring prior information of ships in a large-range sea surface; establishing a GNSS-S radar batch target tracking system model; the method comprises the following steps: establishing a GNSS-S radar batch target tracking optimization model based on a GNSS-S radar batch target tracking problem; and solving the GNSS-S radar batch target tracking optimization model to obtain a beam pointing distribution and beam residence time distribution scheme, and completing GNSS-S radar beam resource optimization. According to the method, the prior information of the search mode is fully utilized, and the target tracking number is greatly increased through optimal allocation of beam resources.

Description

Technical Field

[0001] The present invention relates to the field of radar technology in the electronic information industry, and in particular to a GNSS-S radar beam resource optimization method for on-orbit batch ship tracking. Background Art

[0002] Batch ship target tracking is an important research direction of remote sensing satellite earth observation for marine safety, and plays an important role in maritime traffic, maritime search and rescue, and combat command. The traditional satellite data processing mode is to transmit the data to the ground within a limited transit time after the satellite acquires the data. Due to the limitations of data transmission bandwidth and satellite transit time, the timeliness of target tracking is reduced. To solve the above problems, on-orbit data processing can be used to track ship targets on the satellite and transmit the tracking results to the ground. However, this places high demands on the computing power of the on-board system and the computational complexity of the algorithm.

[0003] Phased array radar systems can form several beams at the same time, and these beams can independently perform tracking and other tasks, so they are widely used in batch target tracking. Due to the agility of the beam, the dwell time of the beam on the target can be controlled by controlling the steering of the beam. Compared with traditional mechanical scanning radars, phased array radars have obvious advantages in the field of multi-target tracking. However, the system resources of the radar are limited, and it is indispensable to propose a reasonable resource allocation plan to complete the batch target tracking task. Summary of the invention

[0004] In order to solve the technical problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a GNSS-S radar beam resource optimization method for on-orbit batch ship tracking, which can adjust the radar beam and dwell time allocation on-orbit to achieve the tracking of a maximum number of ship targets.

[0005] To achieve the above-mentioned object of the invention, the present invention provides a GNSS-S radar beam resource optimization method for on-orbit batch ship tracking, comprising the following steps:

[0006] Step S1, obtaining prior information of ships within a large sea area;

[0007] Step S2, establishing a GNSS-S radar batch target tracking system model;

[0008] Step S3: Based on the GNSS-S radar batch target tracking problem, a GNSS-S radar batch target tracking optimization model is established;

[0009] Step S4, solving the GNSS-S radar batch target tracking optimization model, obtaining the beam pointing allocation and beam dwell time allocation scheme, and completing the GNSS-S radar beam resource optimization.

[0010] According to a technical solution of the present invention, in step S1, a forward squint search mode is used to observe a large range of sea surface, and a low-resolution large-range observation is performed;

[0011] The orbit height of the spaceborne GNSS-S radar system is H. The antenna beam performs forward oblique scanning from near to far. The number of beam scans is M, and the dwell time of the mth beam is T. m , m=1,2,3,…,M, the width corresponding to the mth beam is W m , then the search width of the spaceborne GNSS-S radar system is expressed as:

[0012]

[0013] The prior information includes at least a rough position and RCS of the vessel.

[0014] According to a technical solution of the present invention, the step S2 specifically includes:

[0015] Step S21, based on the radar equation, according to the prior information of each ship and the distance from the spaceborne GNSS-S radar system, calculate the synthetic aperture time required for the phased array radar to meet high-resolution observation;

[0016] Step S22: According to the Kalman filter expression, a state transfer equation and a measurement equation for extracting distance and angle information from the GNSS-S radar echo signal are obtained.

[0017] According to a technical solution of the present invention, in step S3, the beam pointing and dwell time resource management problem is transformed into an optimization problem with tracking the most ship targets as the objective function and the number of beams, total dwell time and tracking accuracy as constraints;

[0018] The BCRLB integrated target distribution, target RCS, process noise and dwell time variable parameters are used to provide the first lower bound for the unbiased estimation of tracking accuracy;

[0019] The first lower bound of the tracking accuracy error is extracted. At tracking time k, when the Bayesian Fisher information matrix at time k–1 is known, given a set of beam variables and the dwell time variable ΔT k =[ΔT 1,k ,…,ΔT Q,k ] T ;

[0020] When the target estimated state is Then the minimum mean square error of the target state estimation satisfies:

[0021]

[0022] in, is the mathematical expectation of the target state and observation; is the BCRLB of the target state estimation error, Indicates the target state The information matrix from Fisher is divided into two parts:

[0023]

[0024] represents the target prior information matrix, is the information matrix of the data,

[0025]

[0026] in, The Jacobian of the measurement function;

[0027] For the target q that is not observed by the beam, the dwell time is zero, and the beam variable is substituted The Fisher information matrix is ​​expressed as:

[0028]

[0029] Taking the inverse, we can get BCRLB as:

[0030]

[0031] The diagonal elements of the target BCRLB represent the variance of position and velocity, which is converted into tracking accuracy error as:

[0032]

[0033] Among them, the diagonal elements of the target BCRLB provide the second lower bound for the minimum root mean square error of the target position estimation, and the second lower bound is used as the criterion of tracking accuracy;

[0034] Based on the constraints, the number of targets that meet a certain tracking accuracy η is maximized, and the algorithm objective function is: δ(·) is a logical function, and the GNSS-S radar batch target tracking optimization model is:

[0035]

[0036]

[0037] Among them, st represents the constraint condition, B' k is the total number of current beams, and B is the upper limit of the number of beams ΔT min , ΔT maxare the minimum and maximum dwell time respectively.

[0038] According to a technical solution of the present invention, in step S4, the optimization problem is converted into a two-step convex optimization solution to form a beam pointing allocation and a beam dwell time allocation scheme, which specifically includes:

[0039] Step S41, judging whether observation is required according to the prior information matrix of each target at the current moment;

[0040] Step S42: combining the coverage of multiple targets by a single beam, a greedy strategy is adopted to achieve coverage of as many targets as possible by a single beam in a two-dimensional plane;

[0041] Step S43: After the beam pointing is assigned, a convex optimization problem of dwell time allocation is formed, and the dwell time allocation result is obtained by using the interior point method.

[0042] According to a technical solution of the present invention, the step S41 specifically includes:

[0043] Based on the target BFIM, the target prior information matrix at the current moment is only related to the information matrix at the previous moment, so the beam is selected based on the target prior information matrix, which is expressed as:

[0044] According to the k-time prior information matrix The inverse is:

[0045] make As a criterion for measuring target prediction and tracking accuracy;

[0046] Will Compared with the tracking threshold η, we get

[0047]

[0048] That is, only the targets whose tracking accuracy does not meet the requirements are observed.

[0049] According to a technical solution of the present invention, the step S42 specifically includes:

[0050] When allocating beams, priority is given to targets with more prior information to achieve more target observations. The target prediction and tracking accuracy vector is expressed as:

[0051]

[0052] Arranged from smallest to largest, they are:

[0053]

[0054] The smaller the value of the tracking accuracy criterion, the better the prior information matrix of the target is, and less observation time is required for state update. k >B, the beam selection method is as follows:

[0055]

[0056] Wherein, M is the number of points that do not need to be irradiated.

[0057] According to a technical solution of the present invention, the step S43 specifically includes:

[0058] Step S431, select the aforementioned targets to be observed, and calculate the beam coverage with them as the center;

[0059] Step S432, assigning a beam center to the area with the largest number of beam overlaps;

[0060] Step S433, removing the covered target;

[0061] Step S434: Repeat steps S432 and S433 until all the targets to be observed are covered, and update Obtaining a beam pointing allocation scheme;

[0062] Step S435: Based on the beam pointing allocation result Converting the original non-convex problem into a convex problem that is easier to solve, we have:

[0063]

[0064]

[0065] Step S436: Obtain the residence time allocation result ΔT by using the interior point method k .

[0066] According to a technical solution of the present invention, the step S21 specifically includes:

[0067] In the short time of imaging, the target in the observation area is considered to be moving at a uniform speed, so the initial position of the ship target q=(1,2,…Q) is The initial speed is Then at time k, the position and velocity of the ship target q are and

[0068] Among them, not all ship targets need to be observed at time k, so the binary variable is introduced in

[0069]

[0070] GNSS-S radar receives GNSS scattered signals, and the duration of a single code signal is T pri , when receiving M for a specific ship target at time k q,k code period, then the observation time of the beam on the ship target is ΔT q,k =M q,k ·T pri ;

[0071] Noise q,k (t,m) is independent of landing power, has a mean of zero, and a variance of Gaussian white noise, according to the radar equation, the echo signal-to-noise ratio is:

[0072]

[0073] Where A is the antenna area, α q,k is the channel loss, σ q,k is the target RCS, P av Indicates the landing power of the GNSS scattered signal;

[0074] Echo signal-to-noise ratio SNR q,k It is proportional to the observation time. Given the observation time, the corresponding echo signal-to-noise ratio SNR can be obtained. q,k .

[0075] According to a technical solution of the present invention, the step S22 specifically includes:

[0076] The position and velocity of the ship target q at time k are The state equation of the Kalman filter is:

[0077]

[0078] Among them, F k represents the state transfer matrix, process noise The mean is 0 and the variance is Q q,k-1 Gaussian white noise, state transfer matrix F k and variance Q q,k-1 They are:

[0079]

[0080]

[0081] Among them, I2 is the second-order identity matrix, represents the Kronecker product operation of the matrix, T s represents the beam sampling interval, κ q represents the process noise intensity of the target q;

[0082] At time k, the target measurement equation of the Kalman filter is expressed as:

[0083]

[0084] Among them, the measurement noise v q,k The mean is zero and the covariance is G q,k Gaussian white noise.

[0085] Compared with the prior art, the present invention has the following beneficial effects:

[0086] The present invention proposes a GNSS-S radar beam resource optimization method for on-orbit batch ship tracking. Through an algorithm with linear computational complexity, the radar beam and dwell time allocation can be adjusted on-orbit to achieve the tracking of a maximum number of ship targets, providing an efficient technical means for ship statistics and detection of illegal ships going out to sea. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0088] Figure 1 A flowchart schematically showing a method for optimizing GNSS-S radar beam resources for on-orbit batch ship tracking according to an embodiment of the present invention;

[0089] Figure 2 The overall flow chart of the GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to one embodiment of the present invention is schematically shown;

[0090] Figure 3 A detection scene diagram of a GNSS-S radar system according to an embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0091] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0092] like Figures 1 to 3As shown, a GNSS-S radar beam resource optimization method for on-orbit batch ship tracking of the present invention comprises the following steps:

[0093] Step S1, obtaining prior information of ships within a large sea area.

[0094] The detection scenario of the GNSS-S radar system is shown in the figure below. Figure 3 As shown in the figure, the forward squint search mode is used to observe a large range of sea surface and perform low-resolution large-scale observation. The orbit height of the spaceborne GNSS-S radar system is H, the antenna beam performs forward squint scanning from near to far, the number of beam scans is M, and the dwell time of the mth beam is T m , m=1,2,3,…,M, the width corresponding to the mth beam is W m , then the search width of the spaceborne GNSS-S radar system is expressed as:

[0095]

[0096] The prior information includes at least a rough position and RCS of the vessel.

[0097] Step S2: establishing a GNSS-S radar batch target tracking system model, specifically including:

[0098] Step S21, based on the radar equation, according to the prior information of each ship and the distance from the spaceborne GNSS-S radar system, calculate the synthetic aperture time required for the phased array radar to meet high-resolution observation;

[0099] Step S22: According to the Kalman filter expression, a state transfer equation and a measurement equation for extracting distance and angle information from the GNSS-S radar echo signal are obtained.

[0100] GNSS-S radar can synthesize B beams simultaneously in the observable area. There are Q targets distributed in the area, and generally B≤Q. In the short time of imaging, the ship targets in the observation area are regarded as moving at a uniform speed. The initial position of the ship target q=(1,2,…Q) is The initial speed is At time k, the position and velocity of the ship target q are and Among them, not all ship targets need to be observed at time k, so the binary variable in

[0101]

[0102] GNSS-S radar receives GNSS scattered signals, and the duration of a single code signal is T pri When receiving M for a specific ship target at time k q,kcode period, then the observation time of the beam on the ship target is ΔT q,k =M q,k ·T pri .

[0103] If the GNSS landing power is P av , noise w q,k (t,m) is independent of landing power, has a mean of zero, and a variance of Gaussian white noise, according to the radar equation, the echo signal-to-noise ratio is:

[0104]

[0105] Where A is the antenna area, α q,k is the channel loss, σ q,k is the target RCS.

[0106] When other conditions are the same, the echo signal-to-noise ratio (SNR) q,k Proportional to the observation time, given the observation time, the corresponding SNR can be obtained q,k .

[0107] The position and velocity of the ship target q at time k are expressed as The state equation of the Kalman filter is:

[0108]

[0109] Among them, F k represents the state transfer matrix, process noise The mean is 0 and the variance is Q q,k-1 Gaussian white noise, state transfer matrix F k and variance Q q,k-1 They are:

[0110]

[0111]

[0112] Among them, I2 is the second-order identity matrix, represents the Kronecker product operation of the matrix, T s represents the beam sampling interval, κ q represents the process noise intensity of the target q.

[0113] At time k, the target measurement equation of the Kalman filter is expressed as:

[0114]

[0115] Among them, the measurement noise v q,k The mean is zero and the covariance is G q,kGaussian white noise.

[0116] Step S3: Based on the GNSS-S radar batch target tracking problem, a GNSS-S radar batch target tracking optimization model is established.

[0117] The present invention transforms the beam pointing and dwell time resource management problem into an optimization problem with tracking the most ship targets as the objective function and the number of beams, total dwell time and tracking accuracy as constraints.

[0118] The Bayesian Cramer-Rao lower bound (BCRLB) is used to provide a first lower bound for the unbiased estimation of tracking accuracy by combining the target distribution, target RCS, process noise and dwell time variable parameters. Therefore, the lower bound of the tracking accuracy error is extracted. At tracking time k, when the Bayesian Fisher Information Matrix (BFIM) at time k–1 is known, given a set of beam variables and the dwell time variable ΔT k =[ΔT 1,k ,…,ΔT Q,k ] T , the corresponding BCRLB can be calculated.

[0119] When the target estimated state is The minimum mean square error of target state estimation satisfies:

[0120]

[0121] in, is the mathematical expectation of the target state and observation; is the BCRLB of the target state estimation error, Indicates the target state The BFIM, which can be divided into two parts:

[0122]

[0123] represents the target prior information matrix, is the information matrix (FIM) of the data,

[0124]

[0125] in, Represents the Jacobian of the measurement function.

[0126] For the target q that is not observed by the beam, the dwell time is zero, and the beam variable is substituted BFIM is expressed as:

[0127]

[0128] Taking the inverse, we can get BCRLB as:

[0129]

[0130] The diagonal elements of the target BCRLB represent the variance of position and velocity, which is converted into tracking accuracy error:

[0131]

[0132] A lower bound is provided for the minimum root mean square error of target position estimation and used as a criterion for tracking accuracy.

[0133] The purpose of this invention is to maximize the number of targets that meet a certain tracking accuracy η based on the constraints. The algorithm objective function is: δ(·) is a logical function, so the optimization problem can be expressed as:

[0134]

[0135]

[0136] Among them, st represents the constraint condition, B' k is the total number of current beams, and B is the upper limit of the number of beams ΔT min , ΔT max are the minimum and maximum dwell time respectively.

[0137] Step S4, solving the GNSS-S radar batch target tracking optimization model, converting the optimization problem into a two-step convex optimization solution, forming beam pointing allocation and beam dwell time allocation schemes respectively, and completing the GNSS-S radar beam resource optimization, specifically including:

[0138] Step S41, judging whether observation is required according to the prior information matrix of each target at the current moment;

[0139] Based on the target BFIM, the target prior information matrix at the current moment is only related to the information matrix at the previous moment, so the beam is selected based on the target prior information matrix, which is expressed as:

[0140] According to the k-time prior information matrix The inverse is:

[0141] make As a criterion for measuring target prediction and tracking accuracy;

[0142] Will Compared with the tracking threshold η, we get

[0143]

[0144] That is, only the targets whose tracking accuracy does not meet the requirements are observed.

[0145] Step S42: combining the coverage of multiple targets by a single beam, a greedy strategy is adopted to achieve coverage of as many targets as possible by a single beam in a two-dimensional plane;

[0146] When allocating beams, priority is given to targets with more prior information to achieve more target observations. The target prediction and tracking accuracy vector is expressed as:

[0147]

[0148] Arranged from smallest to largest, they are:

[0149]

[0150] The smaller the value of the tracking accuracy criterion, the better the prior information matrix of the target is, and less observation time is required for state update. k When >B, the beam selection method is as follows:

[0151]

[0152] Wherein, M is the number of points that do not need to be irradiated.

[0153] Step S43: After the beam pointing is assigned, a convex optimization problem of dwell time allocation is formed. The dwell time allocation result is obtained by using the interior point method, which specifically includes:

[0154] Step S431, select the aforementioned targets to be observed, and calculate the beam coverage with them as the center;

[0155] Step S432, assigning a beam center to the area with the largest number of beam overlaps;

[0156] Step S433, removing the covered target;

[0157] Step S434: Repeat steps S432 and S433 until all the targets to be observed are covered, and update Obtaining a beam pointing allocation scheme;

[0158] Step S435: Based on the beam pointing allocation result Converting the original non-convex problem into a convex problem that is easier to solve, we have:

[0159]

[0160]

[0161] Step S436: Obtain the residence time allocation result ΔT by using the interior point methodk .

[0162] Since the initial value of BFIM is positive semidefinite, L(ΔT q,k ) is convex, the objective function and other constraints are linear, so the optimization problem is a convex optimization problem. With the help of the interior point method, the residence time allocation result ΔT can be obtained. k .

[0163] The present invention discloses a GNSS-S radar beam resource optimization method for on-orbit batch ship tracking. The method adopts a phased array radar system. First, the position and RCS information of ships in a large range of sea surface are obtained by using a forward oblique search mode. Benefiting from the flexible beam pointing capability of the phased array, based on the prior information obtained by the search mode, according to the GNSS-S radar parameters and the target position, a two-step convex optimization algorithm is used to achieve the optimal allocation of the phased array radar beam, that is, firstly, the beam pointing allocation is obtained to achieve the minimum beam occupancy for batch target coverage; then, according to the target RCS and imaging geometry, the dwell time of each radar beam is optimized and allocated to achieve the tracking of the maximum number of ship targets.

[0164] Compared with the traditional observation method that uses the target as the beam pointing center and uniform beam dwell time, this method makes full use of the prior information of the search pattern and greatly increases the number of target tracking by optimizing the allocation of beam resources, providing an efficient technical means for ship statistics and detection of illegal ships going out to sea.

[0165] Parts of the present invention that are not described in detail belong to the well-known technology in the art.

[0166] It should be noted that although the embodiments of the present invention described above are illustrative, they are not intended to limit the present invention, and therefore the present invention is not limited to the above specific embodiments. Without departing from the principles of the present invention, any other embodiments obtained by those skilled in the art under the guidance of the present invention are deemed to be within the protection of the present invention.

Claims

1. A GNSS-S radar beam resource optimization method for on-orbit batch ship tracking, characterized in that: The following steps are involved: Step S1, obtaining prior information of ships within a large sea area; Step S2, establishing a GNSS-S radar batch target tracking system model; Step S3: Based on the GNSS-S radar batch target tracking problem, a GNSS-S radar batch target tracking optimization model is established; Step S4, solving the GNSS-S radar batch target tracking optimization model, obtaining the beam pointing allocation and beam dwell time allocation scheme, and completing the GNSS-S radar beam resource optimization.

2. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 1 is characterized in that: In the step S1, a forward oblique search mode is used to observe a large area of ​​the sea surface, and a low-resolution large-area observation is performed; The orbit height of the spaceborne GNSS-S radar system is H. The antenna beam performs forward oblique scanning from near to far. The number of beam scans is M, and the dwell time of the mth beam is T. m , m=1,2,3,…,M, the width corresponding to the mth beam is W m , then the search width of the spaceborne GNSS-S radar system is expressed as: The prior information includes at least a rough position and RCS of the vessel.

3. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 2 is characterized in that: The step S2 specifically includes: Step S21, based on the radar equation, according to the prior information of each ship and the distance from the spaceborne GNSS-S radar system, calculate the synthetic aperture time required for the phased array radar to meet high-resolution observation; Step S22: According to the Kalman filter expression, a state transfer equation and a measurement equation for extracting distance and angle information from the GNSS-S radar echo signal are obtained.

4. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 3 is characterized in that: In step S3, the beam pointing and dwell time resource management problem is transformed into an optimization problem with tracking the most ship targets as the objective function and the number of beams, total dwell time and tracking accuracy as constraints; The BCRLB integrated target distribution, target RCS, process noise and dwell time variable parameters are used to provide the first lower bound for the unbiased estimation of tracking accuracy; The first lower bound of the tracking accuracy error is extracted. At tracking time k, when the Bayesian Fisher information matrix at time k–1 is known, given a set of beam variables and the dwell time variable ΔT k =[ΔT 1,k ,…,ΔT Q,k ] T ; When the target estimated state is Then the minimum mean square error of the target state estimation satisfies: in, is the mathematical expectation of the target state and observation; is the BCRLB of the target state estimation error, Indicates the target state The information matrix from Fisher is divided into two parts: represents the information matrix of the target prior, is the information matrix of the data, in, The Jacobian of the measurement function; For the target q that is not observed by the beam, the dwell time is zero, and the beam variable is substituted The Fisher information matrix is ​​expressed as: Taking the inverse, we can get BCRLB as: The diagonal elements of the target BCRLB represent the variance of position and velocity, which is converted into tracking accuracy error as: Among them, the diagonal elements of the target BCRLB provide the second lower bound for the minimum root mean square error of the target position estimation, and the second lower bound is used as the criterion of tracking accuracy; Based on the constraints, the number of targets that meet a certain tracking accuracy η is maximized, and the algorithm objective function is: δ(·) is a logical function, and the GNSS-S radar batch target tracking optimization model is: Among them, st represents the constraint condition, B' k is the total number of current beams, and B is the upper limit of the number of beams ΔT min , ΔT max are the minimum and maximum dwell time respectively.

5. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 4 is characterized in that: In step S4, the optimization problem is converted into a two-step convex optimization solution to form beam pointing allocation and beam dwell time allocation schemes, which specifically include: Step S41, judging whether observation is required according to the prior information matrix of each target at the current moment; Step S42: combining the coverage of multiple targets by a single beam, a greedy strategy is adopted to achieve coverage of as many targets as possible by a single beam in a two-dimensional plane; Step S43: After the beam pointing is assigned, a convex optimization problem of dwell time allocation is formed, and the dwell time allocation result is obtained by using the interior point method.

6. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 5 is characterized in that: The step S41 specifically includes: Based on the target BFIM, the target prior information matrix at the current moment is only related to the information matrix at the previous moment, so the beam is selected based on the target prior information matrix, which is expressed as: According to the k-time prior information matrix The inverse is: make As a criterion for measuring target prediction and tracking accuracy; Will Compared with the tracking threshold η, we get That is, only the targets whose tracking accuracy does not meet the requirements are observed.

7. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 5, characterized in that: The step S42 specifically includes: When allocating beams, priority is given to targets with more prior information to achieve more target observations. The target prediction and tracking accuracy vector is expressed as: Arranged from smallest to largest, they are: The smaller the value of the tracking accuracy criterion, the better the prior information matrix of the target is, and less observation time is required for state update. k When >B, the beam selection method is as follows: Wherein, M is the number of points that do not need to be irradiated.

8. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 5, characterized in that: The step S43 specifically includes: Step S431, select the aforementioned targets to be observed, and calculate the beam coverage with them as the center; Step S432, assigning a beam center to the area with the largest number of beam overlaps; Step S433, removing the covered target; Step S434: Repeat steps S432 and S433 until all the targets to be observed are covered, and update Obtaining a beam pointing allocation scheme; Step S435: Based on the beam pointing allocation result Converting the original non-convex problem into a convex problem that is easier to solve, we have: Step S436: Obtain the residence time allocation result ΔT by using the interior point method k .

9. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 3, characterized in that: The step S21 specifically includes: In the short time of imaging, the target in the observation area is considered to be moving at a uniform speed, so the initial position of the ship target q=(1,2,…Q) is The initial speed is Then at time k, the position and velocity of the ship target q are and Among them, not all ship targets need to be observed at time k, so the binary variable is introduced in GNSS-S radar receives GNSS scattered signals, and the duration of a single code signal is T pri , when receiving M for a specific ship target at time k q,k code period, then the observation time of the beam on the ship target is ΔT q,k =M q,k ·T pri ; Noise q,k (t,m) is independent of landing power, has a mean of zero, and a variance of Gaussian white noise, according to the radar equation, the echo signal-to-noise ratio is: Where A is the antenna area, α q,k is the channel loss, σ q,k is the target RCS, P av Indicates the landing power of the GNSS scattered signal; Echo signal-to-noise ratio SNR q,k It is proportional to the observation time. Given the observation time, the corresponding echo signal-to-noise ratio SNR can be obtained. q,k .

10. The GNSS-S radar beam resource optimization method for on-orbit batch ship tracking according to claim 3, characterized in that: The step S22 specifically includes: The position and velocity of the ship target q at time k are The state equation of the Kalman filter is: Among them, F k represents the state transfer matrix, process noise The mean is 0 and the variance is Q q,k-1 Gaussian white noise, state transfer matrix F k and variance Q q,k-1 They are: Among them, I2 is the second-order identity matrix, represents the Kronecker product operation of the matrix, T s represents the beam sampling interval, κ q represents the process noise intensity of the target q; At time k, the target measurement equation of the Kalman filter is expressed as: Among them, the measurement noise v q,k The mean is zero and the covariance is G q,k Gaussian white noise.

Images