46 results about "Transfer orbit" patented technology

The present invention provides a method and system for optimizing the electric propulsion orbit transfer strategy for low-Earth orbit satellite constellation networking, including: Step 1: Based on the engineering constraints of only two states of on and off according to the thrust magnitude of the electric propulsion, establish an optimization problem of the time-optimal constellation networking electric propulsion transfer orbit; Step 2: Divide the networking process into stages, and perform phased parameterization on the thrust magnitude and direction; Step 3: Determine the variables to be solved, and convert the optimization problem of the constellation networking electric propulsion transfer orbit into a shooting problem with constrained control variables; Step 4: Guess the initial value of the shooting problem, and use the Newton iteration algorithm to solve it, and finally obtain the optimal solution of the optimization problem of the constellation networking electric propulsion transfer orbit. The present invention can ensure that the transfer orbit meets the preset networking constraints, has low requirements for the attitude control system, has a simple structure, high solution efficiency, is conducive to on-board implementation, and can effectively realize the joint control of orbital elements.

The invention provides a low-energy transfer window rapid design method and system for earth-moon DRO, and relates to the technical field of aerospace. Based on a double-circle restrictive four-body model, an earth-moon transfer orbit database is constructed, the earth-moon transfer orbit database comprises a plurality of transfer orbits from a target near-earth orbit to a target long-distance retrograde orbit; determining a target track point from the track distribution diagram corresponding to each track leveraging category according to the track information track distribution diagram of each transfer track and the input near-earth track information and epoch time of the target spacecraft, and correcting the track information of the transfer track corresponding to the target track point based on an ephemeris model, and obtaining a target transfer orbit of the target spacecraft. According to the method provided by the invention, rapid and accurate planning of the earth-moon DRO low-energy transfer window can be realized, the processing time is reduced, and the planning efficiency is improved.

The invention discloses a construction method and system for a low-thrust transfer orbit family between earth-moon translation point orbits, and belongs to the technical field of aerospace. The method comprises the following steps: firstly, solving a time optimal transfer orbit based on a continuous low-thrust spacecraft motion equation; then taking the orbit as a seed, performing forward and reverse phase continuation along a departure orbit and an arrival orbit by adopting a self-adaptive step length numerical continuation method, and generating a time optimal transfer orbit family covering different phases; secondly, mapping the time optimal orbit family into an energy optimal orbit family based on control law consistency, and expanding the transfer time range of the energy optimal orbit family through time continuation; and finally, gradually reducing homotopy parameters by using a homotopy method, detecting and processing thrust switching discontinuity in combination with a switching function, and converting the energy optimal orbit family into a fuel optimal transfer orbit family. According to the method, a continuous and complete orbit family is efficiently generated from a single solution, and systematicness and engineering applicability of orbit design are improved.

The invention provides an earth-moon transfer orbit iterative calculation method based on earth-moon relative motion, which comprises the following steps: firstly, calculating an ideal earth-moon transfer orbit number according to the deviation between the current flight time and the theoretical time, and taking the ideal earth-moon transfer orbit number as an iterative guidance target orbit; and then, by utilizing the characteristic that the shutdown time of the rocket in the flight section can be predicted in the iterative guidance calculation process, the influence of the flight time deviation on the earth-moon transfer orbit is compensated to the number of target orbits in real time, so that correction of a guidance program angle instruction is realized, and the rocket is controlled to fly towards a more ideal earth-moon transfer orbit all the time.

The invention discloses a spacecraft approaching rendezvous guidance method based on limited thrust, and belongs to the technical field of spacecraft guidance. The invention aims to solve the problem that fuel consumption and hardware cost cannot be taken into consideration at the same time in an existing spacecraft short-distance guide section autonomous approaching method. Comprising the steps of calculating a theoretical speed increment of a current task stage; in the execution process of the task stage, the state of the propeller is judged according to a preset period; if the propeller is in the propelling process, the remaining speed increment is calculated by combining the theoretical speed increment and the single-cycle speed increment of the propeller; determining that the propeller is in a non-propelling state until the current residual speed increment is smaller than a set threshold value; the theoretical speed increment is calculated again and compared with a track safety belt threshold value to judge whether the tracking spacecraft exceeds the safety belt range or not, and if yes, the task state is updated, and the propeller is controlled to conduct propelling so as to correct the transfer orbit until the task stage is finished. According to the invention, high-precision guidance control is realized.

The application discloses a transfer orbit design method and system of a lunar orbit large ellipse frozen orbit based on eccentricity vector targeting. The transfer orbit lifting type is determined in advance in a transfer stage, the transfer orbit is selected as a lifting orbit, launching and arriving are lifted, the lunar orbit is ensured as a large ellipse frozen orbit with a far point above the south hemisphere of the moon, the eccentricity vector targeting is used to generate a transfer orbit meeting task constraints, the near point amplitude angle is adjusted near the far point after the moon is captured, the lunar orbit large ellipse frozen orbit is efficiently generated, and finally, a pulse is applied at the near point according to task requirements to transfer to a large ellipse frozen orbit with a required period.

The invention provides a battle system double-encircling detection method and system, and the method comprises the steps: carrying out the emission in a form of a combination of a polar orbit encircling device and a medium-low inclination angle encircling device, jointly completing the flight from the earth to a battle transfer orbit, carrying out the selective separation of the two encircling devices before approaching a battle, carrying out the capturing, braking and ignition, and enabling the two encircling devices to enter different orbits, the two surrounding devices are each provided with a magnetic field measurement device, an inter-device communication device and a device-ground communication device, the two surrounding devices carry out star magnetic field and gravitational field measurement respectively, a machine is selected to carry out star atmosphere occultation and plasma ring measurement, and when resources are sufficient, the middle-low dip angle surrounding device can adjust a track to carry out ring wood and satellite detection. The star system double-surrounding detection system comprises the links of transfer orbit design, assembly resource analysis, star system inner orbit design, resource analysis, load addition and distribution demonstration, environment-wood-satellite extended task analysis and the like. According to the invention, two surrounding devices can be launched at a time to obtain the star data, the resource utilization rate and the result output are improved, and the engineering value is high.

The invention discloses an earth-moon L1 point track crossing behavior prediction method and system based on machine learning. According to the method, firstly, an initial state set is generated based on an earth-moon plane circular restrictive three-body model, and an orbit sample data set is constructed through numerical integration; secondly, a multi-class behavior classification system is established, and sample data is automatically labeled; on the basis, an XGBoost algorithm is adopted for model training, a category self-adaptive threshold optimization strategy based on a verification set is introduced for the problem of data extreme imbalance, and the overall prediction precision and category fairness of the model are remarkably improved by independently optimizing an optimal decision threshold for each category and applying the optimal decision threshold to normalized probabilistic decision. According to the method, the technical blank of spacecraft behavior prediction in the earth-moon translation point area is filled, and an effective means is provided for space situation awareness.

The invention discloses a method and system for dynamically reconstructing a launch window for a lunar orbit, and relates to the technical field of lunar exploration launch orbit design. The method comprises the following steps: receiving the initial transfer orbit elements of the lunar orbit, and calculating the orbit elements at the moment of satellite-rocket separation according to the initial transfer orbit elements of the lunar orbit; inputting the orbit elements at the moment of satellite-rocket separation into an accurate dynamics equation, iteratively obtaining the accurate orbit elements of the first lunar orbit at the take-off moment, changing the take-off moment, and iteratively obtaining the accurate orbit elements of multiple lunar orbits; performing polynomial fitting on the bound elements of the accurate orbit elements of the multiple lunar orbits to obtain the polynomial coefficients corresponding to each orbit, and binding the elements of a launch vehicle as the polynomial coefficients corresponding to each orbit, so as to realize dynamic reconstruction of the lunar orbit according to the bound elements of the launch vehicle.

The present invention discloses a method for designing an orbit of a kinetic impact mission on a near-Earth asteroid with the optimal deflection effect, comprising the following steps: S1, selecting the initial parameters of the spacecraft, including the departure and arrival times of the spacecraft; S2, designing the spacecraft orbit as a splicing of three segments, namely a large elliptical Earth parking orbit, an Earth escape orbit starting from the large elliptical orbit, and an interplanetary transfer orbit, and calculating the orbit parameters; S3, evaluating the deflection effect of the target near-Earth asteroid; S4, traversing all possible departure and arrival times of the spacecraft, repeating steps S1 to S3, drawing a contour map of the deflection distance versus the departure and arrival times of the spacecraft, and determining the optimal spacecraft launch window from it after eliminating unreasonable times. By adopting a large elliptical Earth parking orbit and screening the spacecraft launch window, the present invention enables the mass and relative velocity of the spacecraft to be relatively large when impacting the asteroid, thereby achieving the optimal asteroid deflection effect.

The application provides a spacecraft thruster fault response method, comprising: transferring a small geostationary orbit satellite from an earth transfer orbit to a geostationary orbit through three stages; the small geostationary orbit satellite comprises an orbit transfer chemical thruster, an attitude control chemical thruster and an electric thruster, wherein, in the second stage of the transfer of the small geostationary orbit satellite, a fault response scheme is started according to a fault response strategy of "first electric propulsion, then orbit transfer chemical thruster, and then attitude control chemical thruster", and the fault response scheme comprises: when the electric thruster is faulty, the orbit transfer chemical thruster is started; when both the electric thruster and the orbit transfer chemical thruster are faulty, the attitude control chemical thruster is started.

The invention particularly relates to a method and a system for estimating the minimum velocity increment of multi-circle double-pulse orbit transfer under J2 perturbation. The method comprises the following steps: uniformly representing origin-destination orbit states as six spring equinox elements taking flat longitude as a phase quantity; aiming at a transfer structure of a first pulse, middle drift and a second pulse, determining a limited transfer turn number search window by utilizing a flat longitude drift characteristic and a pulse capability; a closed equality constraint system with the total variable quantity = the first pulse variable quantity + the middle-section drift variable quantity + the second pulse variable quantity is constructed, and it is ensured that the track state is strictly closed; pulse is decomposed into tangential, normal and radial components for cooperative solution, and a phase closed loop is realized by updating a semi-diameter increment; and performing low-dimensional nonlinear optimization in a limited number of turns to quickly obtain minimum speed increment estimation. According to the method, the calculation efficiency is remarkably improved while the precision is guaranteed, and the method is suitable for large-scale sequence optimization in the near-earth orbit multi-target access task.

The present invention discloses a method and system for optimizing the design of Earth-Moon orbits under a four-pulse ephemeris model based on manifold splicing, and relates to the technical field of designing probe transfers between special Earth-Moon orbits. The technical key points of the present invention include: processing the initial values ​​of the Halo orbits corresponding to the Sun-Earth system and the Earth-Moon system based on invariant manifolds, and obtaining the optimal manifold splicing points of the Halo orbits corresponding to the Sun-Earth system and the Earth-Moon system respectively; constructing a four-pulse dual-circle restrictive four-body model; using the optimal manifold splicing points of the Halo orbits corresponding to the Sun-Earth system and the Earth-Moon system as initial values, and using a particle swarm optimization algorithm and a sequential quadratic programming algorithm to solve the four-pulse dual-circle restrictive four-body model to obtain the optimal spliced ​​orbit. The present invention uses two optimization models to design low-energy transfer orbits, which can accelerate iterations, improve computational efficiency and algorithm convergence, and realize rapid search for low-energy transfer orbits in specific launch windows, meeting the needs of space stations for Earth-Moon low-energy orbit transfers.

The launch vehicle plus upper stage joint orbit design method comprises the following steps: step 1, taking the launch vehicle as a base stage and the upper stage as a terminal stage, a three-degree-of-freedom joint trajectory optimization model is established; step 2, according to the target task characteristics, the joint design orbit section is determined; the joint design orbit section comprises the upper stage working times and the transfer orbit type; step 3, the upper stage first working starting point is determined, and the joint optimization is completed; after the transfer orbit is determined, the launch vehicle base stage and the upper stage first working section are subjected to joint orbit optimization, and the upper stage first working starting point is determined. The launch vehicle plus upper stage joint orbit design method can greatly improve the carrying capacity of the combination of the launch vehicle base stage and the upper stage by jointly optimizing the trajectories of the launch vehicle base stage and the upper stage and determining the upper stage first working starting point, and has high application value in engineering.

The invention discloses a real-time two-dimensional calculation method for a ground fire transfer orbit based on a variable step size method, and the method mainly comprises the following steps: 1, setting flight time, and obtaining gravitational constants of the earth, the Mars and the sun; 2, initializing the system; 3, judging a time stage; 4, screening time steps; 5, calculating the motion trail of the spacecraft based on the Newton's law of motion and the universal gravitation law; 6, solving the position and speed of the spacecraft through a numerical integration method; 7, updating the mass change of the spacecraft in real time; and 8, drawing a spacecraft orbit and outputting related data. The method can be used for simulating the orbit behavior of the spacecraft between the earth and the Mars, especially under the condition that a nuclear power propulsion system is considered. According to the method, the motion trail of the spacecraft in the complex gravitational field can be accurately calculated based on classical mechanics and numerical simulation technologies, and theoretical support is provided for design and optimization of aerospace tasks.

The invention discloses a lunar exploration mission rocket trajectory and spaceship orbit combined reconstruction method, which comprises the following steps: S1, after a thrust drop fault occurs in a three-level flight section of a carrier rocket, judging whether to continue to execute a lunar flight mission or not; s2, if the execution is continued, determining an initial equivalent departure moment and an initial transfer orbit parameter under the current fault condition based on an orbit perigee equivalent conversion method; s3, obtaining a new transfer orbit parameter satisfying the matching of the lunar orbit constraint and the earth-moon transfer time; s4, solving to obtain a reconstructed trajectory of the rocket and a corresponding actual orbit injection point state and an actual orbit injection moment; and S5, calculating the transfer time from the orbit perigee corresponding to the actual orbit injection point state to the actual orbit injection point. According to the method, after the thrust drop fault of the carrier rocket occurs, the rocket flight trajectory and the airship earth-moon transfer orbit can be quickly and synergistically reconstructed online, so that the airship can still enter the earth-moon orbit according to task requirements under the fault condition.

The present disclosure provides a spacecraft orbit transfer control method, device and storage medium suitable for deep space ferry mission, belonging to the field of spacecraft guidance and design. The method comprises: obtaining a dynamic model of a first spacecraft transferring from a first orbit to a second orbit; based on the dynamic model, an initial transfer orbit is obtained through a first homotopy parameter with energy optimization as the target; based on the initial transfer orbit, a target transfer orbit is obtained through a second homotopy parameter and a third homotopy parameter with fuel optimization as the target. The present disclosure provides an initial value guess for the fuel optimization problem by using the easily solvable energy optimization problem through the first homotopy parameter; through the synergistic effect of the second homotopy parameter and the third homotopy parameter, the target transfer orbit under the fuel optimization target is obtained, thereby reducing the sensitivity of the indirect method to the initial value and enhancing the convergence of the solution of the complex orbit transfer task.

The present application relates to a method for designing a transfer orbit for a HALO orbiter to fly over a given ground point. The method includes: finding the moment when the lunar declination value is equal to the latitude value of the given ground point within the given mission time range, and selecting the time point closest to the initial moment as the initial value of the iteration for the orbit transfer moment; calculating the flight duration of the orbit transfer point relative to the starting point using the starting time and the initial value of the iteration for the orbit transfer moment; calculating the normalized time by dividing the flight duration by the HALO orbit period; calculating the magnitude of the orbit transfer velocity increment based on the normalized time using the polynomial fitting coefficients; establishing a mapping relationship between the magnitude of the orbit transfer velocity increment and the initial value of the iteration for the orbit transfer moment through numerical integration based on a high-precision force model; and completing the solution of the parameter to be solved using the Newton iteration method based on the mapping relationship. The method adopted has a short response time, high calculation result accuracy, and fast iteration convergence.

The application discloses a kind of multi-satellite optimal orbit optimization method in planetary system inside borrowing force fuel consumption, belong to aerospace technical field.The application implementation method is: selecting the sequence of borrowing force satellite satisfying detection demand, with each transfer orbit period as variable, only adding maneuver when leaving initial orbit and being captured or entering target orbit by target satellite, the optimization target is minimum total velocity increment, and the orbit optimization problem is constructed under planetary mass center inertia system.According to the state of spacecraft in the parameter solving of transfer orbit between borrowing force satellites, the state of next satellite borrowing force is solved, the parameter of transfer orbit after borrowing force can be matched from the state to realize continuous solving, until the state when reaching target satellite is obtained.The optimization problem is solved using global optimization algorithm, the relative velocity of reaching each borrowing force satellite and the transfer orbit parameters before and after borrowing force are obtained, and the transfer of each segment transfer orbit under reference ephemeris is realized according to the transfer orbit parameters.

The application discloses a transfer orbit parameter optimization selection method of a lunar-circulating large-ellipse frozen orbit and belongs to the technical field of deep space exploration orbits. The method comprises the following steps: setting parameters that can be optimized in a transfer process of a lunar-circulating large-ellipse frozen orbit; analyzing influences of different near-lunar-point altitudes, inclination angles and periods of a capture orbit on velocity increments, measurement and control conditions and orbit safety; analyzing influences of different near-lunar-point altitudes, inclination angles and near-lunar-point amplitude angles of a transition orbit on velocity increments, measurement and control conditions and orbit safety; and analyzing influences of different mission orbit periods on velocity increments and orbit safety. If all are optimal, the optimization of the transfer orbit parameters is ended; if not, a mission orbit period is selected by means of bias parameter optimization. The application can obtain a complete capture-to-target orbit scheme and efficiently design orbit parameters from the moon capture to the large-ellipse frozen orbit.

The invention discloses an earth-moon transfer orbit accurate control method and system, and the method comprises the steps: introducing different types and sizes of deviations in the initial state and flight process of a detector through a Monte Carlo simulation method, simulating the orbit propagation process of the detector under the influence of the deviations, and constructing a deviation propagation vector model; according to the deviation propagation vector model, deviation values of different flight stages are predicted, and the optimal correction opportunity and correction value are determined in combination with track measurement data of the detector; and based on the optimal correction opportunity and the correction amount, in combination with a pre-constructed multi-objective optimization model, obtaining an optimal control scheme, and based on the optimal control scheme, controlling the earth-moon transfer orbit. Wherein the multi-objective optimization model is constructed by taking the number of times of correction, the opportunity of each correction and the correction amount as decision variables and taking the total fuel consumption amount as a constraint condition, so that the orbit control precision of the detector can be effectively improved, the on-orbit life can be prolonged, and reliable orbit control support can be provided for a deep space exploration task.

Embodiments of the present application disclose a transfer orbit design method from GTO to the Sun-Earth collinear libration point, a device and a medium; the method comprises: calculating and obtaining an invariant manifold connected with the Halo orbit around the libration point according to the spacecraft dynamics equation and the potential function in the Sun-Earth barycenter coordinate system; taking the geosynchronous transfer orbit as the earth parking orbit to construct the constraint variable, taking the Halo orbit as the target orbit to construct the control variable, and correcting the control variable by using the least square differential correction method; based on the set constraint condition and performance index, the optimal HOI point corresponding to the optimal performance transfer orbit is selected from the transfer orbit meeting the constraint condition; according to the optimal HOI point, the relationship between the control variable and the parking orbit constraint variable is analyzed, and the differential correction initial value expression for determining the control variable initial value of the multi-constraint orbit design is constructed; the transfer orbit meeting multiple constraints is obtained by using the multi-stage differential correction method according to the differential correction initial value expression.

The invention discloses a transfer orbit design method in on-orbit service of a synchronous orbit satellite. The problem of how to effectively image a high-orbit target satellite with low fuel consumption is solved. Belongs to the technical field of space. Comprising the following steps: determining total transfer time according to different initial states of a service satellite and a target satellite; constructing a transfer orbit mathematical model based on the total transfer time to obtain the running time of the service satellite on the initial orbit, the transfer orbit and the approaching orbit in the transfer process and the speed increment for completing the Horman transfer; taking a relative position vector included angle and an illumination angle of two stars at the end of intersection as an objective function, taking total transfer time as a constraint condition, taking the operation time of a service star in an initial orbit and an approaching orbit as an optimization variable, optimizing the objective function by using a genetic algorithm based on an HPOP model, and correcting parameters of a transfer orbit mathematical model to obtain a transfer orbit mathematical model; obtaining a transfer orbit mathematical optimization model; according to the method, effective imaging of the high-orbit target with low burnup is realized through the mathematical optimization model of the transfer orbit.

The present invention provides a rapid design method and system for the transfer orbit of the Sun-Earth L5 libration point. Based on the heliocentric two-body orbit analytical model, the number of transfer loops, transfer time, and eccentricity of the transfer orbit are set, and the semi-major axis of the transfer orbit is calculated. According to the semi-major axis, eccentricity of the transfer orbit in step 1 and the orbital parameters of the Earth, the inclination of the transfer orbit is set, and the Earth zero-radius sphere of influence model is used to calculate the magnitude and direction of the hyperbolic excess velocity of the detector. It is set that the detector applies a braking impulse once when arriving at the L5 point, and the magnitude and direction of the impulse are calculated by using the transfer orbit parameters in steps 1-2. Taking the heliocentric two-body transfer orbit parameters calculated by the above steps as the initial value guess, the heliocentric multi-body orbit model is used to calculate the high-precision transfer orbit through iterative optimization. The method in the present invention realizes the trade-off among orbit accuracy, computational amount, and design efficiency, effectively reduces the number of iterations, reduces the computational amount, and ensures the robustness of numerical iteration.

The application provides a small-thrust spacecraft trajectory optimization method and system based on a hybrid expert model, and relates to the technical field of spacecraft trajectory optimization. The application designs and implements an intelligent fusion prediction architecture based on a gating network and a hybrid expert model. The architecture solves the problems of insufficient generalization ability and precision decline of existing single models when facing diversified tasks. Each expert network is trained as a characteristic network for processing a specific type of small-thrust trajectory transfer trajectory. In a multi-target rendezvous task scenario, the hybrid expert model can autonomously fuse the prediction results of each expert, and show high-stable and high-precision prediction results on any type of small-thrust transfer orbit. The robustness and generalization performance of the system prediction are far superior to those of existing single models.

The present application relates to a method for designing a transfer orbit from a DRO orbit to a low lunar circular orbit. The method includes: establishing a mapping data table and a three-dimensional data pair of the delta-v of orbit transfer according to the preset DRO orbit size parameters and the phase angle of the DRO vehicle at the initial moment; dividing the three-dimensional data pair of the delta-v of orbit transfer into two groups according to the positive and negative conditions of the high and low angles of the delta-v of orbit transfer, obtaining the right ascension of the ascending node and the transfer flight time corresponding to the data pair with the minimum delta-v during the transfer process in each group to calculate the orbit transfer moment, and selecting the minimum value of the two orbit transfer moments as the selected orbit transfer moment; correcting the phase of the selected orbit transfer moment, and calculating the projection of the delta-v of orbit transfer according to the corrected phase; using the Newton iteration method to perform iterative solution with the selected orbit transfer moment and the projection of the delta-v of orbit transfer as the initial iteration values to obtain the transfer orbit transfer moment and the transfer delta-v of orbit transfer. Using this method can improve the design accuracy of the transfer orbit.

The application discloses a transfer orbit electric propulsion satellite thrust drop detection method and system, comprising: establishing a single electric thruster real-time thrust-force variation relation model according to satellite data; calculating a satellite force tolerance range according to the single electric thruster real-time thrust-force variation relation model based on the single electric thruster real-time thrust-force variation relation model; if the electric thruster has a fault, determining the installation position of the faulty electric thruster first; according to the force size at different time within the satellite working time range and the installation position of the faulty electric thruster, determining the fault mode of the faulty electric thruster by using the fault mode function of the faulty electric thruster, and generating a faulty thruster thrust curve. The application can effectively track any form of thrust variation, lock the faulty thruster and determine the fault mode, so that timely correction measures are taken to ensure that the orbit and attitude of the satellite are kept in the predetermined target range.

A method and system for orbital transfer under the Earth-Moon three-body model based on a deep neural network relates to the technical field of probe transfer design between special Earth-Moon orbits, and is used to solve the technical problem that existing methods cannot solve due to large computational complexity and long time consumption. The technical key points of the present invention include: obtaining a training dataset for transfer between special orbits; constructing a deep neural network and training the deep neural network based on the special transfer training dataset to obtain a trained deep neural network; constructing an optimization model for direct transfer between special orbits of the Earth-Moon system, and iteratively solving the direct transfer optimization model using a particle swarm algorithm and the trained deep neural network until the desired index or the maximum number of iterations is reached, thereby obtaining an iterative index curve and a transfer orbit under the Earth-Moon three-body model. The present invention realizes the calculation and design of optimal transfer pulses between special orbits of the Earth-Moon system.

The invention provides a low-thrust spacecraft trajectory optimization method and system based on a hybrid expert model, and relates to the technical field of spacecraft trajectory optimization. According to the invention, a set of intelligent fusion prediction architecture based on the gating network and the hybrid expert model is designed and realized. The architecture solves the problems that the generalization ability is insufficient and the precision is reduced when an existing single model faces diversified tasks. Each expert network is trained as a characteristic network for processing a special type of low-thrust trajectory transfer trajectory. In a scene requiring a multi-target rendezvous task, the hybrid expert model can autonomously fuse prediction results of all experts, a high-stability and high-precision prediction result is shown on any type of low-thrust transfer orbit, and the robustness and generalization performance of system prediction are far better than those of an existing single model.

The invention relates to a transfer orbit design method for a HALO orbital vehicle to fly to a given ground point. The method comprises the following steps: finding a moment when a lunar declination value is equal to a given ground point latitude value according to the lunar declination value in a given task time range, and selecting a time point closest to an initial moment as an orbit transfer moment iteration initial value; calculating the flight duration of the orbital transfer point relative to the starting point by using the starting point time and the iteration initial value of the orbital transfer time; calculating normalization time according to the quotient of the flight duration and the HALO orbital period; according to a polynomial fitting coefficient, calculating the increment of the orbital transfer speed for the normalization time; based on a high-precision force model, establishing a mapping relation between the increment of the orbital transfer speed and the iteration initial value at the orbital transfer time through numerical integration; and solving the to-be-solved parameters by adopting a Newton iteration method based on the mapping relation. The method is short in response time, high in calculation result precision and fast in iteration convergence.