Dynamic target relay observation task planning method based on trajectory prediction

Through the method based on trajectory prediction and neural network, combined with the contract network algorithm, the problem that a single satellite cannot be continuously observed by a high-speed aircraft during high-speed flight is solved, and effective relay observation and coverage of dynamic targets are achieved.

CN119090168APending Publication Date: 2024-12-06BEIJING RES INST OF TELEMETRY
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202410972061.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-19
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

It is difficult for the prior art to conduct continuous observation of high-speed aircraft, especially when the target flight speed is fast and the infrared signal is relatively weak, a single satellite cannot conduct continuous observation effectively, and there is a lack of real-time information on the target's movement during relay observation to guide.

Method used

The dynamic target relay observation task planning method based on trajectory prediction is adopted, and the target motion trajectory is predicted through mathematical modeling and neural networks, and the contract network algorithm is combined with the contract network algorithm to quickly allocate the task to the next satellite to achieve relay observation of the dynamic target.

Benefits of technology

Continuous observation of high-speed aircraft is achieved, satellite switching frequency is reduced, and the coverage and efficiency of relay observation are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119090168A_ABST
    Figure CN119090168A_ABST
Patent Text Reader

Abstract

The invention provides a trajectory prediction-based dynamic target relay observation task planning method. The method comprises the following steps of: firstly, performing mathematical modeling on dynamic target relay observation task planning; and according to observed target motion information, predicting a moving track of the target after the target is separated from a satellite view field, generating a relay observation task of a next stage according to a track prediction result and a current state of a satellite constellation, and quickly distributing the task to a next satellite by adopting a contract net algorithm to realize relay observation of the dynamic target. According to the invention, the relay observation of the dynamic target is realized, the problem that the relay observation satellite lacks task information guidance when the satellite carries out relay observation on the dynamic target is solved, and the satellite switching frequency in the relay observation process is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of measurement and testing technology, and in particular to a dynamic target relay observation task planning method based on trajectory prediction. Background Art

[0002] In the existing space-based reconnaissance system, the tracking of high-speed flying targets mainly relies on monitoring by high-orbit infrared satellites. Such aircraft usually use rocket engines as power, and have obvious infrared characteristics during flight. Monitoring satellites can monitor large areas in high orbits.

[0003] However, the rapidly developing near-space high-speed aircraft in recent years often use rocket-ramjet combination engines with smaller infrared characteristics or adopt boost-glide mode for flight. In the initial stage when the rocket engine is working, it can be detected by high-orbit satellites. After entering the ramjet engine flight phase or the glide flight phase, the infrared characteristics are greatly weakened. At this time, the target cannot be detected and tracked by high-orbit satellites.

[0004] When using low-orbit satellites that are closer to the target for detection, although low-orbit satellites have the characteristics of high detection sensitivity, due to their low orbit altitude and small satellite field of view, they are unable to continuously track the target.

[0005] Therefore, when using a large-scale low-orbit constellation to continuously observe high-speed targets in near-space, it is necessary to call different satellites to relay tracking of the targets. In the existing relay tracking task planning methods, most of them are aimed at low-speed targets or targets that can be continuously monitored by high-orbit satellites. When conducting relay observations on low-speed targets, the main consideration is the changes in the field of view caused by the movement of the satellite itself, and the target can be approximately regarded as not moving. When conducting relay observations on targets that can be continuously monitored by high-orbit satellites, low-orbit satellites mainly play the role of conducting detailed inspections under the guidance of high-orbit satellites, and the information obtained from high-orbit satellite observations can be used to guide low-orbit satellites in real time. High-speed targets in near-space have the characteristics of fast movement speed and lack of guidance information. The target cannot be regarded as stationary, and there is a lack of real-time information on the target's movement for guidance during relay observations.

[0006] Therefore, it is necessary to adopt a multi-satellite coordinated relay observation method for continuous observation, but there is currently no suitable solution. Summary of the invention

[0007] The present invention aims to solve the problem that the target has a high flying speed and a relatively weak infrared signal, and a high-speed aircraft cannot be effectively and continuously observed by a single satellite. A dynamic target relay observation task planning method based on trajectory prediction is provided. First, mathematical modeling is performed on the dynamic target relay observation task planning. According to the observed target motion information, the running trajectory of the target after leaving the satellite field of view is predicted, and the relay observation task of the next stage is generated according to the trajectory prediction result and the current state of the satellite constellation. The contract network algorithm is used to quickly assign the task to the next satellite to realize the relay observation of the dynamic target.

[0008] The present invention provides a dynamic target relay observation task planning method based on trajectory prediction, comprising the following steps:

[0009] S1. Detection by high-orbit satellites during the initial rocket engine operation phase of the target spacecraft;

[0010] S2. After the target spacecraft enters the ramjet flight phase or the gliding flight phase, the high-orbit satellite provides the low-orbit constellation with the target initial information x0. The target initial information x0 can guide the first low-orbit satellite to start observing the target. The target initial information x0 is the initial information of the target observation task set X, x0 = {τ0, p0, t0}, where τ0 is the start time of the initial observation task, p0 is the initial target position, t0 is the initial observation duration, t0 = 0; the i-th low-orbit satellite conducts the i-th observation of the target spacecraft, i = 1;

[0011] S3, the current low-orbit satellite calculates the target motion trajectory information based on the observation data and predicts the target motion trajectory after it leaves the current low-orbit satellite observation range through the neural network-based observation target trajectory prediction method to obtain the predicted post-deorbit observation information x' i+1 , x i+1 '={τ' i+1 ,p' i+1 ,t i+1}, where τ i+1 ' is the delay time of the i+1th observation task relative to the start time τ0 of the initial observation task, p i+1 'Predict the target location for the i+1th observation task, t i+1 is the duration of the i+1th observation mission, t i+1 Updated at the end of the i+1th observation mission;

[0012] S4, the current low-orbit satellite will contain the predicted post-deorbit observation information x' before the target spacecraft leaves the observation range i+1 The relay observation tasks are published to the task sequence to be planned to obtain the prediction target observation task set X', X' = {x0, ..., x i ,xi+i ',…,x n};

[0013] S5. After receiving the predicted target observation task set X', other satellites calculate whether the satellite and the observed target are visible according to the predicted trajectory in the task information, and calculate the visible observation window of the observed target;

[0014] S6. Other satellites calculate their own bid values ​​through window information. The satellite that has just completed the observation mission and released the relay observation mission starts the contract network bidding process to collect the bidding information of other satellites. The bid value calculated by each satellite is:

[0015] Q j =a·τ j +b·t j ,

[0016] Among them, a and b are weight coefficients, τ j is the start time of the observation mission of the j-th bidding satellite to the observation target, t j is the observation mission duration of the j-th bidding satellite on the observation target;

[0017] S7, the bidding satellite selects the satellite with the best bid value as the winning satellite, and sends a message to confirm the allocation result;

[0018] The criterion for determining the best bid value is that the low-orbit constellation has the highest observation coverage of the target's operation process;

[0019] S8. After receiving the confirmation information, the winning satellite starts relay observation at the scheduled time;

[0020] The winning satellite determines whether i+1 is equal to n. If so, a dynamic target relay observation mission planning method based on trajectory prediction is completed. If not, i+1=i, and returns to step S3.

[0021] The dynamic target relay observation task planning method based on trajectory prediction described in the present invention is, as a preferred embodiment,

[0022] In step S2, X={x0,…,x i ,x i+1 ,…,x n}, 0<i<n, where x i is the information of the relay observation task of the i-th segment, x i ={τ i ,p i ,t i}, τ i is the delay time of the i-th relay observation task relative to the start time τ0 of the initial observation task; p i is the position of the target at the beginning of the i-th relay observation task; ti is the duration of the i-th relay observation task.

[0023] The dynamic target relay observation task planning method based on trajectory prediction described in the present invention is, as a preferred embodiment, in step S3, the observation target trajectory prediction method based on a neural network includes the following steps:

[0024] S3A, offline training: First, the state estimation value and the state truth value of the target at any adjacent time in the aircraft trajectory library are preprocessed to generate input-output pairs suitable for network training; then, the regularized square error loss function is used to train the deep network, which includes a deep long short-term memory network;

[0025] S3B, online prediction: For the observed target trajectory at the current sampling time T, the target state estimate at time T+1 is estimated using the same preprocessing method as offline training; then the target state prediction value x output by the deep network is T+1|1:T As input, it is sent to the deep long short-term memory network again to obtain the state prediction value x at time T+2 T+2|1:T ; After multiple recursive forecasts, the long-period trajectory forecast of the active segment of the observed target is obtained.

[0026] The present invention discloses a dynamic target relay observation task planning method based on trajectory prediction, as a preferred embodiment, in step S3A, a measurement sequence 1: K-1 is obtained from an aircraft trajectory library and output to a tracker, the tracker performs trajectory estimation, estimates the trajectory according to the measurement sequence 1: K-1 to obtain a state estimation value, the measurement sequence 1: K-1 is a state true value, the state estimation value and the state true value are paired for preprocessing to obtain an input-output pair and output to a deep network, a true trajectory sequence 2: K is obtained from the aircraft trajectory library, the measurement sequence 1: K-1 and the true trajectory sequence 2: K are adjacent time data, the true trajectory sequence 2: K is preprocessed and the deep network is trained using a regularized square error loss function.

[0027] The dynamic target relay observation mission planning method based on trajectory prediction described in the present invention is, as a preferred mode, the data source of the aircraft trajectory library is part of the key target flight data obtained through monitoring, reconnaissance and intelligence collection.

[0028] The dynamic target relay observation task planning method based on trajectory prediction described in the present invention is, as a preferred embodiment, the data in the aircraft trajectory library is the flight data of the target aircraft generated by utilizing a dynamic model and typical values ​​of target flight parameters.

[0029] The present invention discloses a dynamic target relay observation task planning method based on trajectory prediction. As a preferred method, the data in the aircraft trajectory library is to achieve target flight parameter traversal by deviating the nominal value of the target flight parameter, solve the dynamic model, and generate trajectory data that meets the time and altitude.

[0030] The dynamic target relay observation task planning method based on trajectory prediction described in the present invention is, as a preferred embodiment, the deep network includes three long short-term memory network layers, a maximum output layer and a fully connected layer, wherein the number of neurons in the three long short-term memory network layers are selected as 128, 256, and 256 respectively, the number of neurons in the maximum output layer is selected as 64, and the number of neurons in the fully connected layer is selected as 6.

[0031] The dynamic target relay observation task planning method based on trajectory prediction of the present invention is a preferred method. In step S5, the angle and distance between the target and the satellite line and the satellite-ground line are calculated to determine whether the target is within the visible range of the satellite. The determination method is arccos (r s ·(r s -r t ) / ||r s ||·||r t ||)≤θ,||r s ||≤h·cos(θ), where r s is the position vector of the satellite in the geocentric inertial coordinate system, r t is the position vector of the target in the geocentric inertial coordinate system, θ is the half field of view of the satellite, and h is the altitude of the satellite.

[0032] The dynamic target relay observation task planning method based on trajectory prediction described in the present invention is, as a preferred embodiment, in step S6, the length of each satellite observation window in the execution of the task can be adjusted by setting a and b to adjust the number of satellites that complete the task;

[0033] In step S7, the criterion for determining the optimal bid value is to maximize the sum of the tracking durations for the dynamic targets.

[0034] Since a single satellite is effectively and continuously observing a high-speed aircraft, when planning the relay observation mission, it is necessary to: first fit the target flight trajectory and predict the trajectory based on the observation data of the previous satellite, generate the relay observation task based on the trajectory prediction result, and finally assign the relay observation task to the satellite that meets the constraints for execution.

[0035] The present invention has the following advantages:

[0036] (1) The present invention designs a method for generating and planning relay observation tasks based on trajectory prediction. The method can predict the trajectory of the target after it leaves the satellite's field of view based on the observed target motion information through a neural network method, generate the next stage of relay observation tasks based on the trajectory prediction results and the current state of the satellite constellation, and use the contract network algorithm to quickly assign the task to the next satellite, thereby realizing relay observation of dynamic targets.

[0037] (2) The present invention overcomes the problem that when the satellites conduct relay observation of dynamic targets, the relay observation satellites lack mission information guidance;

[0038] (3) The present invention reduces the frequency of satellite switching during the relay observation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a flow chart of a dynamic target relay observation task planning method based on trajectory prediction;

[0040] Figure 2 This is a trajectory prediction scheme flow chart for a dynamic target relay observation task planning method based on trajectory prediction. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0042] Example 1

[0043] A dynamic target relay observation task planning method based on trajectory prediction includes the following steps:

[0044] (1) Detection by high-orbit satellites during the initial rocket engine operation phase of the target spacecraft;

[0045] (2) After the target spacecraft enters the ramjet flight phase or the gliding flight phase, the high-orbit satellite provides the low-orbit constellation with the target initial information x0 to guide the first low-orbit satellite to start observing the target;

[0046] x0∈X={x0,x1,…,x i-1 ,x i ,x i+1 ,…,x n}, where x i ={τ i ,p i ,t i} represents task related information. τ i represents the delay time of the i-th relay observation task relative to the start time τ0 of the initial observation task; pi represents the location of the target at the beginning of the i-th relay observation task; t i represents the duration of the i-th relay observation;

[0047] (3) The low-orbit satellite that receives the guidance information observes the target according to the guidance information, calculates the target's motion trajectory information based on the observation data, and then predicts the target's motion trajectory after it leaves the satellite's observation range through a deep network model;

[0048] The trajectory prediction method includes two stages: offline training and online prediction. In the offline training stage, the state estimation value and the true state value of the target at any adjacent time are first preprocessed to generate input-output pairs suitable for network training; then, the designed deep long short-term memory network is trained using the regularized square error loss function. In the online prediction stage, for the observed target trajectory at the current sampling time T, the target state estimation value at time T+1 is estimated using the same preprocessing method; then, the target state prediction value x output by the network is converted into the target state prediction value x. T+1|1:T As input, it is sent to the network again to obtain the state prediction value x at time T+2 T+2|1:T ; By analogy, the active long-period trajectory prediction of the observed target can be realized;

[0049] (4) before the target leaves the current satellite observation range, the relay observation task containing the target predicted trajectory information is released to the task sequence to be planned;

[0050] X={x0,x1,…,x s ,…,x n}, where x i ={τ i ,p i ,t i} represents task related information. When i>s, the time of the future relay observation task is τ i With position p i The information is generated by the trajectory prediction algorithm and is s Update before ending; t i The information is updated at the end of each relay observation mission;

[0051] (5) After other satellites receive the published relay observation mission, they calculate whether the satellite and the mission target are visible based on the predicted trajectory in the mission information, and calculate the visible observation window of the target;

[0052] By calculating the angle and distance between the target and the satellite line and the satellite-ground line, we can determine whether the target is within the visible range of the satellite. The judgment method is arccos (r s ·(r s -r t ) / ||r s||·||r t ||)≤θ,||r s ||≤h·cos(θ), where r s is the position vector of the satellite in the geocentric inertial coordinate system, r t is the position vector of the target in the geocentric inertial coordinate system, θ is the half field of view angle of the satellite, and h is the altitude of the satellite;

[0053] (6) Calculate its own bid value through window information. The satellite that has just completed the observation mission and released the relay observation mission will start the contract network bidding process to collect the satellite's bidding information; each satellite calculates the bid value Q j =a·τ j +b·t j , where a and b are weight coefficients. The length of each satellite observation window in the mission can be adjusted by setting a and b to adjust the number of satellites that complete the mission;

[0054] (7) After collecting the satellite bidding information, the bidding satellite selects the satellite with the best bidding value as the winning satellite and sends a message to confirm the allocation result;

[0055] The principle of large-scale constellation dynamic target relay observation mission planning is to maximize the sum of tracking time for dynamic targets, that is, the constellation's observation coverage of the target's operation process is the highest;

[0056] (8) After receiving the confirmation information, the winning satellite will start relay observation at the scheduled time, and continue to predict the target's future flight trajectory based on the trajectory information obtained from the observation, and then guide the subsequent satellites to relay observation;

[0057] There are two sources of trajectory data: the first is to obtain some key target flight data through surveillance, reconnaissance, intelligence collection and other means; the second is to generate the flight data of the target aircraft using the dynamic model and the typical values ​​of the target flight parameters. In this embodiment, the target flight parameters are traversed by pulling the nominal values ​​of the target flight parameters, solving the dynamic model, and generating 12,000 trajectories that meet the time and altitude.

[0058] The deep network includes three long short-term memory network layers, a maximum output layer and a fully connected layer. The number of neurons in the three long short-term memory network layers is selected as 128, 256, and 256 respectively, the number of neurons in the maximum output layer is selected as 64, and the number of neurons in the fully connected layer is selected as 6;

[0059] The observed target trajectory prediction method based on neural network includes two stages: offline training and online prediction. Figure 2As shown. In the offline training stage, the state estimation value and the true state value of the target at any adjacent time are first preprocessed to generate input-output pairs suitable for network training; then, the designed deep long short-term memory network is trained using the regularized square error loss function. In the online prediction stage, for the observed target trajectory at the current sampling time T, the target state estimation value at time T+1 is estimated using the same preprocessing method; then, the target state prediction value x output by the network is converted into T+1|1:T As input, it is sent to the network again to obtain the state prediction value x at time T+2 T+2|1:T ; By analogy, the active long-period trajectory prediction of the observed target can be realized.

[0060] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A dynamic target relay observation task planning method based on trajectory prediction, characterized in that: The following steps are involved: S1. Detection by high-orbit satellites during the initial rocket engine operation phase of the target spacecraft; S2. After the target aircraft enters the ramjet flight phase or the gliding flight phase, the high-orbit satellite provides the low-orbit constellation with the target initial information x0. The target initial information x0 can guide the first low-orbit satellite to start observing the target. The target initial information x0 is the initial information of the target observation task set X, x0 = {τ0, p0, t0}, where τ0 is the start time of the initial observation task, p0 is the initial target position, t0 is the initial observation duration, t0 = 0; the i-th low-orbit satellite performs the i-th observation of the target aircraft, i = 1; S3, the current low-orbit satellite calculates the target motion trajectory information based on the observation data and predicts the target motion trajectory after it leaves the current low-orbit satellite observation range through the neural network-based observation target trajectory prediction method to obtain the predicted post-deorbit observation information x' i+1 , x i+1 '={τ' i+1 ,p' i+1 ,t i+1 }, where τ i+1 ' is the delay time of the i+1th observation task relative to the start time τ0 of the initial observation task, p i+1 'Predict the target location for the i+1th observation task, t i+1 is the duration of the i+1th observation mission, t i+1 Updated at the end of the i+1th observation mission; S4. The current low-orbit satellite will contain the predicted post-deorbit observation information x' before the target spacecraft leaves the observation range. i+1 The relay observation tasks are published to the task sequence to be planned to obtain the prediction target observation task set X', X' = {x0, ..., x i ,x i+i ',…,x n }; S5. After receiving the predicted target observation task set X', other satellites calculate whether the satellite and the observed target are visible according to the predicted trajectory in the task information, and calculate the visible observation window of the observed target; S6. Other satellites calculate their own bid values ​​through window information. The satellite that has just completed the observation mission and released the relay observation mission starts the contract network bidding process to collect the bidding information of other satellites. The bid value calculated by each satellite is: Q j =a·τ j +b·t j , Among them, a and b are weight coefficients, τ j is the start time of the observation mission of the j-th bidding satellite to the observation target, t j is the observation mission duration of the j-th bidding satellite on the observation target; S7, the bidding satellite selects the satellite with the best bid value as the winning satellite, and sends a message to confirm the allocation result; The criterion for determining the best bid value is that the low-orbit constellation has the highest observation coverage of the target's operation process; S8, the winning satellite starts relay observation at a predetermined time after receiving the confirmation information; The winning satellite determines whether i+1 is equal to n. If so, a dynamic target relay observation task planning method based on trajectory prediction is completed. If not, i+1=i, and returns to step S3.

2. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 1, characterized in that: In step S2, X={x0,…,x i ,x i+1 ,…,x n }, 0<i<n, where x i is the information of the relay observation task of the i-th segment, x i ={τ i ,p i ,t i }, τ i is the delay time of the i-th relay observation task relative to the start time τ0 of the initial observation task; p i is the position of the target at the beginning of the i-th relay observation task; t i is the duration of the i-th relay observation task.

3. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 1, characterized in that: In step S3, the neural network-based observation target trajectory prediction method comprises the following steps: S3A, offline training: firstly, preprocess the state estimation value and the state true value of the target at any adjacent time in the aircraft trajectory library to generate input-output pairs suitable for network training; then, use the regularized square error loss function to train the deep network, which includes a deep long short-term memory network; S3B, online prediction: For the observed target trajectory at the current sampling time T, the target state estimate at time T+1 is estimated using the same preprocessing method as offline training; then the target state prediction value x output by the deep network is T+1|1:T As input, it is sent to the deep long short-term memory network again to obtain the state prediction value x at time T+2 T+2|1:T ; After multiple recursive forecasts, the long-period trajectory forecast of the active segment of the observed target is obtained.

4. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 3 is characterized in that: In step S3A, a measurement sequence 1: K-1 is obtained from the aircraft trajectory library and output to a tracker, the tracker performs trajectory estimation, estimates the trajectory according to the measurement sequence 1: K-1 to obtain a state estimation value, the measurement sequence 1: K-1 is a state true value, the state estimation value and the state true value are paired for preprocessing to obtain an input-output pair and output to the deep network, a real trajectory sequence 2: K is obtained from the aircraft trajectory library, the measurement sequence 1: K-1 and the real trajectory sequence 2: K are adjacent time data, the real trajectory sequence 2: K is preprocessed and the deep network is trained using the regularized square error loss function.

5. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 3 is characterized in that: The data source of the aircraft trajectory library is some key target flight data obtained through surveillance, reconnaissance and intelligence collection.

6. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 3 is characterized in that: The data in the aircraft trajectory library is flight data of the target aircraft generated by utilizing a dynamics model and typical values ​​of target flight parameters.

7. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 6 is characterized in that: The data in the aircraft trajectory library is obtained by traversing the target flight parameters by deviating the nominal values ​​of the target flight parameters, solving the dynamics model, and generating trajectory data that meets the time and altitude requirements.

8. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 3 is characterized in that: The deep network includes three long short-term memory network layers, a maximum output layer and a fully connected layer, wherein the number of neurons in the three long short-term memory network layers are selected as 128, 256, and 256 respectively, the number of neurons in the maximum output layer is selected as 64, and the number of neurons in the fully connected layer is selected as 6.

9. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 1, characterized in that: In step S5, the angle and distance between the target and the satellite line and the satellite-ground line are calculated to determine whether the target is within the visible range of the satellite. The determination method is arccos (r s ·(r s -r t ) / ||r s ||·||r t ||)≤θ,||r s ||≤h·cos(θ), where r s is the position vector of the satellite in the geocentric inertial coordinate system, r t is the position vector of the target in the geocentric inertial coordinate system, θ is the half field of view of the satellite, and h is the altitude of the satellite.

10. The method for planning a dynamic target relay observation task based on trajectory prediction according to claim 1, characterized in that: In step S6, the length of each satellite observation window in the mission can be adjusted by setting a and b to adjust the number of satellites that complete the mission; In step S7, the criterion for determining the optimal bid value is to maximize the sum of the tracking durations for the dynamic targets.

Citation Information

Cited By

  • Multi-satellite cooperative continuous observation and real-time data unloading method for dynamic target

    CN119727874A

  • Method for forecasting uncertainty trajectory of target by low-orbit star network

    CN121655548A

  • Cross-satellite vehicle target tracking method combining road network and target information

    CN121982068A

  • A cross-satellite vehicle target tracking method combining road network and target information

    CN121982068B