A method for constructing a simulation system for a relative navigation device of an intelligent spacecraft

Abstract

The application discloses a simulation system construction method for a relative navigation device of an intelligent spacecraft, which integrates a spaceflight dynamics calculation module into a simulation engine to construct a unified dynamics data driven framework; a space environment simulation model is established based on celestial body position data, an illumination model and a star table space index mechanism to realize rapid generation and dynamic updating of a star background; a spacecraft simulation model is constructed in combination with a spacecraft structure, material and camera installation pose parameters, and spacecraft movement is driven by dynamics data; through camera physical imaging modeling, noise modeling and a multi-modal data synchronous generation mechanism, image, depth map and point cloud data with time sequence consistency are output. The method solves problems such as insufficient coupling of system dynamics and rendering, difficulty in balancing simulation efficiency and precision of a star background and lack of multi-modal data, and can provide a high-fidelity ground simulation verification environment for development, testing and performance evaluation of an intelligent spacecraft relative navigation algorithm.

Description

Technical Field

[0001] This invention belongs to the field of spacecraft simulation and relative navigation equipment simulation, specifically relating to a method for constructing a simulation system for relative navigation equipment for intelligent spacecraft. Background Technology

[0002] Space relative navigation is a crucial foundational technology for intelligent spacecraft performing missions such as rendezvous and docking, on-orbit servicing, rendezvous and reconnaissance, space debris removal, and approaching non-cooperative targets. Relative navigation equipment typically employs optical sensors, such as visible light cameras, to acquire image information of the target intelligent spacecraft, and combines image processing and pose estimation algorithms to calculate the target's relative position and attitude. Therefore, constructing a high-fidelity, repeatable, and verifiable simulation test environment for relative navigation equipment during the ground phase is of great significance for improving the reliability of navigation algorithms and mission safety.

[0003] Existing relative navigation image simulation methods typically separate attitude and orbit dynamics calculations from graphics rendering. Information exchange between systems occurs via network communication or data interfaces, which can easily lead to timing discrepancies and data synchronization errors between the dynamic state and the image generation process, making it difficult to meet the real-time and timing consistency requirements of complex space missions. In stellar background simulation, existing systems lack a rapid star retrieval and real-time rendering mechanism for dynamic changes in the camera's field of view, making it difficult to balance the real-time performance and realism of stellar background simulation.

[0004] Furthermore, existing simulation technologies mostly focus on spacecraft platform status monitoring or single subsystem modeling, lacking a mechanism for frame-level synchronous output of image, depth map, and point cloud data under a unified time reference, making it difficult to support the development and verification of relative navigation algorithms. Therefore, there is an urgent need for a simulation system construction method for intelligent spacecraft relative navigation equipment to improve the realism, completeness, temporal consistency, and credibility of ground verification. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the prior art, such as insufficient coupling between dynamics and rendering, difficulty in balancing the efficiency and accuracy of stellar background simulation, and lack of multimodal data. It provides a method for constructing a simulation system for relative navigation equipment of intelligent spacecraft, realizing the full-link simulation construction from space dynamic environment, target intelligent spacecraft motion, space environment lighting and stellar background modeling to real visible light camera imaging and navigation algorithm verification.

[0006] To achieve the above technical objectives, the present invention adopts the following solution: a method for constructing a simulation system for relative navigation equipment of intelligent spacecraft, comprising:

[0007] Step 1: Establish a high-precision aerospace dynamics solution module coupled with the simulation engine to solve the spatial relative position parameters of the main celestial bodies in real time under a unified simulation time reference, as well as the orbit and attitude state parameters of the tracking intelligent spacecraft and the target intelligent spacecraft, as the underlying data source of the simulation system.

[0008] Step 2: Based on the spatial relative position parameters of celestial bodies and the principle of optical radiation, construct a space environment illumination model; construct a stellar background model based on the star catalog. The space environment illumination model and the stellar background model together constitute the space environment simulation model.

[0009] Step 3: Drive the intelligent spacecraft model to perform relative motion based on the orbit and attitude state parameters, construct a three-dimensional model of the spacecraft based on the physical structure and material parameters of the spacecraft, and configure the installation parameters of the relative navigation camera to complete the construction of the spacecraft system-level simulation model;

[0010] Step four: Based on the installation parameters and attitude parameters of the relative navigation camera, determine the spatial pose of the relative navigation camera and calculate the spatial line-of-sight vector; configure the physical imaging parameters of the relative navigation camera to establish an adjustable visible light camera imaging model; based on the space environment simulation model, spacecraft simulation model, and visible light camera imaging model, perform physical imaging simulation and noise modeling on the target intelligent spacecraft to generate a continuous image data stream; under a unified simulation time reference, synchronously generate and output visible light images, depth maps, point cloud data, and spacecraft orbital attitude state data, establish the correspondence between the observed data and the true state data, and complete the construction of the relative navigation device-level simulation model.

[0011] Compared with the prior art, the present invention has the following beneficial effects:

[0012] (1) By constructing a system-level simulation system covering attitude and orbit calculation, space environment simulation, target imaging and multimodal data generation, the simulation of the entire process from attitude and orbit state to observation data output is realized, thereby improving the real-time performance, integrity and timing consistency of the relative navigation equipment simulation system.

[0013] (2) By coupling the aerospace dynamics calculation module with the simulation engine using a unified time reference, the timing of dynamic state updates and image rendering processes can be made consistent, thereby reducing data synchronization errors generated during multi-system collaboration.

[0014] (3) By constructing a space environment simulation model based on celestial position data, spatial illumination model and star catalog spatial indexing mechanism, we can realize the rapid query, dynamic update and real-time rendering of star background, improve the efficiency of star background generation while ensuring the simulation authenticity, and take into account both simulation accuracy and real-time performance.

[0015] (4) By constructing a visible light imaging simulation model that includes the spacecraft’s structural materials, surface reflection characteristics and camera physical imaging process, the consistency between the visible light images of the target intelligent spacecraft and the real space observation results is improved, and the simulation realism of the optical imaging process in complex space environments is enhanced.

[0016] (5) Through the multimodal data synchronous generation mechanism under a unified time reference, the collaborative output of visible light images, depth maps, point cloud data and spacecraft orbital attitude state data is realized, and the temporal correspondence between observation data and real state data is established, providing complete and reliable data support for the development, performance evaluation, accuracy analysis and test verification of relative navigation algorithms. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the overall process of constructing a simulation system for relative navigation equipment of intelligent spacecraft according to the present invention.

[0018] Figure 2 A flowchart for constructing a stellar background model based on a star catalog.

[0019] Figure 3 Build a flowchart for point clouds.

[0020] Figure 4 This is a flowchart of the simulation system operation of the present invention.

[0021] Figure 5 This is a module architecture diagram of the simulation system of the present invention. Detailed Implementation

[0022] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0023] A method for constructing a simulation system for relative navigation equipment of intelligent spacecraft, such as Figure 1 As shown, it includes four core steps, and the specific implementation method is as follows:

[0024] Step 1: Establish a high-precision aerospace dynamics solution module coupled with the simulation engine to solve the position parameters of major celestial bodies in real time under a unified simulation time reference, as well as the orbit and attitude state parameters of the tracking intelligent spacecraft and the target intelligent spacecraft, as the underlying data source of the simulation system.

[0025] This step aims to provide spatiotemporally consistent and physically accurate input data for the entire simulation system. Its specific construction process includes:

[0026] Step 1.1: Remove the graphical interface from the open-source aerospace dynamics calculation program and compile it into a dynamic link library;

[0027] The open-source general-purpose space mission analysis software GMAT (General Mission Analysis Tool) was selected as the source for aerospace dynamics calculations. To seamlessly integrate with the simulation engine and improve operational efficiency, all interactive components, such as the graphical user interface, were removed from the original program, retaining only the core algorithms for orbit integration and celestial ephemeris calculation, which were then compiled into a dynamic link library.

[0028] Step 1.2: Based on dynamic link libraries, the aerospace dynamics calculation program is repackaged into a dynamics solution module for use by the simulation system.

[0029] Based on the native interface provided by this dynamic link library, and utilizing the GDExtension extension mechanism of the GODOT simulation engine, an extension module encapsulating the calling logic of the aforementioned dynamic link library was written and compiled into an external extension library that can be directly loaded by the GODOT engine. This external extension library registers a custom dynamics solver node with the GODOT engine, allowing users to directly add and call GMAT's dynamics solver functions in the scene. The dynamics solver node is used to establish a data interaction interface between the simulation system and the dynamics solver module, and provides a unified state query capability to the simulation system. The dynamics solver node receives the current simulation time parameters sent by the simulation system, passes them to the dynamics solver module for state calculation, and returns the celestial state data, orbital state data, and attitude state data at the corresponding time.

[0030] During the simulation system's runtime, the system periodically calls the aforementioned dynamics calculation module according to logical step sizes. Each frame call inputs a unified simulation timestamp, thereby acquiring the spatial relative positions of the main celestial bodies precisely aligned with that moment, as well as the position and attitude data of the tracking intelligent spacecraft and the target intelligent spacecraft. This mechanism achieves spatiotemporal synchronization coupling between high-precision dynamics calculation and simulation graphics rendering. All output data is used as the underlying data source for subsequent steps.

[0031] Step 2: Based on the spatial relative position parameters of the celestial bodies and the principle of optical radiation from Step 1, construct a space environment illumination model; construct a space index structure based on the star catalog, and perform rapid retrieval and dynamic updating of visible stars according to the changes in the field of view of the relative navigation camera to construct a stellar background model, thus completing the construction of the space environment simulation model;

[0032] This step constructs a realistic simulation model of the space environment, covering two main components: the lighting environment and the star background.

[0033] Step 2.1: Construct a spatial environment lighting model.

[0034] First, using the spatial relative position parameters of the Sun, Earth, and Moon that are calculated and transmitted in real time in step one, the irradiation direction vector of the Sun relative to the current observation point, as well as the geometric angles and orientations of the Earth, Moon, and other albedo celestial bodies relative to the observation point, are calculated.

[0035] Subsequently, based strictly on the principle of optical radiative transmission, two illumination components were calculated: the direct solar illumination component, whose irradiance is calculated according to the solar constant and the inverse square law of distance; and the celestial albedo component, which uses the directly illuminated celestial body as the light source, treating the celestial body as a diffuse reflective sphere, and employing the average Bond albedo model for calculation, such as 0.30 for Earth and 0.11 for the Moon, to calculate the secondary irradiance reaching the observation point after reflection from the celestial body. Superimposing these two components constitutes a dynamic and complete space environment illumination model.

[0036] Finally, to achieve visual rendering, operations such as tone mapping are used to convert radiance quantities such as physical irradiance into RGB brightness values ​​that can be processed by computer graphics, and display adaptation conversion is completed.

[0037] Step 2.2: Construct a stellar background model based on the star catalog and combine it with the space environment lighting model to complete the construction of the space environment simulation model, such as... Figure 2 As shown:

[0038] (1) First, a high-precision Tycho-2 star catalog is selected. The original star catalog data is preprocessed: the right ascension, declination, proper motion, apparent magnitude, spectral type and other characteristic parameters of each star are extracted one by one, the data integrity is checked, and the coordinates of all stars are uniformly converted to the J2000.0 celestial coordinate system, and finally the star catalog data with standard format is produced.

[0039] (2) Divide the entire celestial sphere into regular sections and establish a celestial sphere index structure. Divide the entire celestial sphere according to right ascension and declination. The first spatial lattice element, the... The coverage of right ascension and declination of the grid cell is as follows:

[0040] ;

[0041] In the formula, Right ascension of a star For the declination of stars, The total number of lattice elements in the right ascension direction. This represents the total number of lattice cells in the declination direction. For cell index, For right ascension index, This is the declination index.

[0042] All stars in the standardized stellar database are loaded into corresponding spatial cells according to their celestial coordinates, forming a stellar index database organized based on celestial spatial location.

[0043] In order to reduce the repeated traversal calculation of all star data during the simulation process, the celestial sphere spatial index structure is established, and the parameters such as the direction vector and irradiance of stars in each spatial cell are pre-calculated and cached.

[0044] For stars with large motions, their coordinates are corrected to the simulation epoch. :

[0045]

[0046] in, The original right ascension of the star. The original declination of the star, For the right ascension to move on its own, For the declination to move, This represents the number of years from J2000.0. The celestial coordinates of each star are then converted to unit direction vectors in the J2000 inertial coordinate system. :

[0047] ;

[0048] In the formula, This is the stellar unit direction vector in the J2000 inertial coordinate system.

[0049] Simultaneously, the irradiance of each star at the camera's entrance pupil is pre-calculated based on its apparent magnitude.

[0050] ;

[0051] in This is the radiation emitted by the star at the camera's entrance pupil. The reference irradiance for zero magnitude stars in the V-band is... The apparent magnitude of the star is determined, and then the irradiance is further converted into a brightness parameter.

[0052] A pre-calculated and cached star index database can avoid frequent calculations such as magnitude conversion and coordinate transformation during simulation, thereby improving the efficiency of real-time generation of star backgrounds.

[0053] (3) By querying the celestial sphere spatial index structure, quickly filter out the spatial cells that intersect with the current field of view and extract the set of visible stars contained therein.

[0054] During the simulation, the unit vector of the camera optical axis direction is obtained in real time based on the relative navigation camera pose information. and field of view First, the optical axis direction vector is solved inversely to obtain celestial coordinates:

[0055] ;

[0056] in: The right ascension corresponding to the camera's optical axis. The declination corresponding to the camera's optical axis. Let be the unit direction vector of the camera's optical axis in the J2000 inertial coordinate system. Using the angular radius, determine the declination and right ascension ranges covered by the current field of view:

[0057] ;

[0058] in, Right ascension for field of view coverage, For the declination of the field of view, This is a longitude contraction correction term along the declination direction, used to compensate for the compression effect of the actual coverage angle width of the right ascension lattice cells in high-latitude regions. The calculation of the right ascension range needs to handle cross-latitude... / Boundary conditions.

[0059] Map the above celestial sphere coverage area to the lattice index:

[0060] ;

[0061] in: The lower and upper bounds of right ascension for the field of view coverage The lower and upper boundaries of declination are defined by the field of view coverage.

[0062] By reading only the stellar data cached within the cells of the aforementioned index range, the visible star set within the current field of view can be quickly extracted without traversing the entire star database one by one. When the camera attitude changes or the field of view expands, the visible star set filtering process is re-executed to dynamically update the stellar data corresponding to the current field of view.

[0063] (4) Based on the direction vector, brightness parameters, and other parameters of each star in the visible star set, the stars are projected onto the camera imaging plane to generate corresponding star background images. The generated star background images can be updated in real time as the camera observation direction changes, thereby constructing a star background model dynamically coupled with the camera's field of view.

[0064] The space environment lighting model and the star background model have been completed and together form the space environment simulation model.

[0065] Step 3: Drive the intelligent spacecraft model to perform relative motion based on the orbit and attitude state parameters described in Step 1. Construct a three-dimensional model of the spacecraft based on the physical structure and material parameters of the spacecraft, and configure the installation parameters of the relative navigation camera to complete the construction of the spacecraft system-level simulation model.

[0066] This step aims to build a simulation model of the tracking intelligent spacecraft and the target intelligent spacecraft.

[0067] First, relative motion mapping is performed. The orbital and attitude parameters of the tracking and target intelligent spacecraft, output from step one, are converted in real time into coordinate transformation matrices in the world coordinate system. Specifically, a translation matrix is ​​generated from the position vector, and a rotation matrix is ​​generated from the attitude quaternions; these two matrices are combined to form a 4×4 homogeneous transformation matrix. In each update of the simulation scenario, these transformation matrices are applied to the root node of the corresponding intelligent spacecraft's 3D model, dynamically updating its 3D spatial pose, thereby accurately reproducing the relative motion relationship between the spacecraft in virtual space.

[0068] Secondly, highly realistic 3D digital modeling was performed. Based on the physical structure and surface material composition of the tracking and target intelligent spacecraft, the 3D model underwent refined material configuration. Specifically, a physics-based bidirectional reflection distribution function model and the Cook-Torrance micro-element model were used to construct the spacecraft surface reflection model. For the three main types of materials commonly found on spacecraft surfaces, corresponding material parameters were configured: for thermal insulation materials, pale yellow or silver color parameters, extremely low metallicity parameters, high roughness parameters, and normal perturbation parameters characterizing wrinkle details were configured; for solar cells, dark blue or black color parameters, moderately low metallicity parameters, and low roughness parameters were configured; for metallic structural materials, high metallicity parameters and roughness parameters consistent with the processing technology were configured. These material parameters characterize the inherent reflection characteristics, metallic reflection properties, high light scattering characteristics, and local geometric details of the model surface.

[0069] Next, a direct mapping relationship was established between the aforementioned material parameters and the material properties of the selected physical rendering engine. Color parameters were mapped to Albedo, metallicity parameters to Metallic, roughness parameters to Roughness, and normal perturbation parameters to Normal Map. By assigning corresponding materials to each structural region of the spacecraft's 3D model, differentiated rendering of the surface reflection characteristics of different structural components was achieved.

[0070] Finally, using the space environment lighting model built in step two as the sole light source input for the scene, the physical rendering engine performs global lighting calculations on the spacecraft surface based on the configured BRDF model, calculates the reflected light radiation brightness in real time, and generates a rendered image with natural transitions between light and dark, realistic highlight shapes, and clear material texture, thereby completing the reproduction of real optical characteristics.

[0071] In addition, this step requires initial configuration of the position parameters relative to the navigation camera, including the camera's installation position and pointing direction in the coordinate system of the tracking intelligent spacecraft, and converting them into a navigation camera installation matrix, which is then recorded in the model attributes to prepare for the camera pose calculation in step four. At this point, the spacecraft system-level simulation model is complete.

[0072] Step four: Based on the installation parameters of the relative navigation camera and the attitude state parameters of the tracking intelligent spacecraft, determine the spatial pose of the relative navigation camera and calculate the spatial line-of-sight vector; configure the physical imaging parameters of the relative navigation camera and establish an adjustable visible light camera imaging model; based on the space environment simulation model, spacecraft simulation model, and visible light camera imaging model, perform physical imaging simulation and noise modeling on the target intelligent spacecraft to generate a continuous image data stream; under a unified simulation time reference, synchronously generate and output visible light images, depth maps, point cloud data, and spacecraft orbital attitude state data, establish the correspondence between the observed data and the true state data, and complete the construction of the relative navigation device-level simulation model.

[0073] This step is used to realize the conversion process from simulation scenario to observation data of relative navigation equipment, and to build a multimodal observation data generation mechanism under a unified time reference.

[0074] Step 4.1: Establish an adjustable visible light camera imaging model and perform layered physical imaging.

[0075] By combining the spacecraft's attitude state and camera installation parameters, the spatial pose and line-of-sight vector of the camera are calculated. An adjustable imaging model is constructed by configuring the physical parameters of the real camera. Through layered imaging, layer compositing, and noise degradation, a time-series continuous visible light image data stream is generated.

[0076] (1) Based on the position and attitude parameters of the tracking intelligent spacecraft output in step one, construct the pose transformation matrix of the tracking intelligent spacecraft in the world coordinate system; combine the relative navigation camera installation matrix configured in step three, calculate the spatial pose of the relative navigation camera in the world coordinate system. Further, calculate the current camera optical axis direction based on the camera attitude matrix and the forward unit vector of the camera local coordinate system, and obtain the spatial line-of-sight vector.

[0077] (2) Construct the color image acquisition camera and depth camera in the relative navigation camera in the simulation engine. The depth camera and the visible light camera in the relative navigation camera share the same spatial pose, camera intrinsic parameters, focal length parameters, field of view parameters and near and far clipping plane parameters, and perform the imaging process synchronously under a unified time reference, thereby establishing a strict pixel-level correspondence between the visible light image and the depth image.

[0078] Based on the actual design specifications of the relative navigation camera, physical imaging parameters are configured to establish a physical visible light camera imaging model. The configured physical imaging parameters include at least: camera intrinsic parameters, extrinsic parameters, focal length, aperture F-number, exposure time, sensor sensitivity, focusing distance, imaging resolution, and field of view. These parameters are set entirely according to the design specifications of the actual relative navigation camera and can be dynamically adjusted.

[0079] After the camera model for the stellar background is constructed, a layered physical imaging process based on unified camera parameters is executed. First, the system uses a color image acquisition camera to render the stellar background model constructed in step two, generating a stellar background image. Then, using the same camera parameters, it renders the target intelligent spacecraft, generating a visible light image of the target intelligent spacecraft. Simultaneously, the depth camera outputs a depth image corresponding to the current scene and records the depth information of each pixel on the target surface. Since the background image and the target image are generated using the same projection model, they satisfy a consistent spatial projection relationship.

[0080] Subsequently, the background image of the star and the image of the target intelligent spacecraft are composited. During the composite process, occlusion determination is performed based on the depth information corresponding to the target image. When the target intelligent spacecraft is in front of the background star, the corresponding pixels of the background image are covered by the pixels of the target image; otherwise, the background star information is preserved, thereby achieving a realistic occlusion effect of the target intelligent spacecraft on the star background and generating a noise-free original visible light image.

[0081] Based on this, noise modeling and degradation processing are performed on the original visible light image. The noise model includes photon shot noise and readout noise, where photon shot noise follows a Poisson distribution, and readout noise includes dark current noise and circuit thermal noise. Simultaneously, a motion blur model is constructed based on the image plane displacement corresponding to the target during the exposure time, and the length and direction of the blur kernel are determined. By superimposing the above time-varying noise, an image data stream with temporal continuity is generated.

[0082] Step 4.2: Generate a 3D point cloud data stream by coupling the depth image from the depth camera in the relative navigation camera with the visible light image from the visible light camera. The process can be found in [link to relevant documentation]. Figure 3 .

[0083] In order to further extract target spatial structure information from scene rendering results, this embodiment further constructs a joint generation mechanism of depth image and 3D point cloud based on continuous image data generation.

[0084] During each frame of scene rendering, the depth camera rendering process automatically generates corresponding depth buffer data, where each pixel records the depth information of the target surface relative to the virtual camera. Since the values ​​in the rendering depth buffer are non-linear depth expressions in normalized device coordinate space, they cannot be directly used as the target's true spatial distance. Therefore, the system first reads the depth buffer and, combined with the virtual camera's projection parameters and the inverse projection matrix, performs spatial reconstruction operations on the depth values ​​to obtain the actual depth value of each pixel's corresponding target surface point in the camera's view space. For any pixel, the depth reconstruction process can be described as follows: first, determine its normalized device coordinates based on the pixel position; then, recover the corresponding view space coordinates through inverse projection transformation; and extract the spatial distance information of the target point along the camera's line of sight, thereby constructing depth image data with floating-point precision.

[0085] Furthermore, the system uses each effective pixel in the depth image as a spatial unit, fusing the color information of the corresponding pixel in the visible light image with the depth value of the corresponding pixel in the depth image. It then uses a virtual camera projection model to perform spatial back-projection calculations to recover the 3D spatial position of the target surface point corresponding to that pixel in the camera coordinate system. The spatial point in the camera coordinate system can be further transformed to the target body coordinate system, spacecraft coordinate system, or scene unified reference coordinate system based on the pose relationship between the virtual camera and the target object, to meet the unified coordinate expression requirements of different navigation scenarios.

[0086] To improve point cloud generation efficiency, the system employs a parallel computing approach using a graphics processor. It synchronously performs spatial back-projection calculations on all pixels in the depth image, using effective pixels as the basic parallel computing unit while simultaneously recovering the corresponding 3D coordinates and color information, thus forming dense 3D point cloud data for the current frame. Compared to traditional CPU-based pixel-by-pixel calculations, this method significantly reduces point cloud generation overhead in high-resolution scenes and improves real-time processing capabilities under continuous image streaming conditions.

[0087] Step 4.3: Unify the simulation time base, synchronously generate and output visible light images, depth maps, point cloud data, and spacecraft orbital attitude state data, establish the correspondence between observed data and true state data, and complete the construction of the relative navigation equipment-level simulation model.

[0088] To ensure time consistency between the dynamics calculation process and the graphics rendering process, a unified simulation time management mechanism is established within the simulation system. This unified simulation time management mechanism maintains a global simulation time variable and updates it incrementally according to logical step sizes. At the beginning of each simulation cycle, the unified simulation time management mechanism generates a simulation timestamp for the current moment and simultaneously distributes it to the dynamics calculation module, space environment simulation module, spacecraft simulation module, and camera imaging module. Each module performs state updates and image generation processes based on the same timestamp.

[0089] While generating visible light images, depth images, and 3D point cloud data, the system simultaneously reads the orbital state parameters and attitude state parameters output from step one and stores them using a unified simulation timestamp. The orbital state parameters include at least a position vector and a velocity vector, and the attitude state parameters include at least attitude quaternions and an attitude matrix. This establishes a one-to-one correspondence between the observed data and the true state data.

[0090] Finally, visible light images, depth images, 3D point cloud data, and spacecraft orbital attitude state data are synchronously output under a unified simulation time base, forming a unified temporal multimodal observation data stream. This multimodal observation data stream can be used as input to the relative navigation algorithm, and combined with the synchronously output state truth data, accuracy verification and performance evaluation are carried out. This completes the construction of the relative navigation device-level simulation model.

[0091] The simulation system for relative navigation equipment of intelligent spacecraft of this invention executes a closed-loop operation process according to a unified simulation time base, such as... Figure 4 As shown, the specific implementation method is as follows:

[0092] The system first acquires the relative spatial positions of major celestial bodies and the orbital and attitude parameters of the tracking intelligent spacecraft and the target intelligent spacecraft based on a high-precision aerospace dynamics calculation module, and outputs temporal pose information. Then, it combines the space environment lighting model, the stellar background model, and the 3D model of the target intelligent spacecraft to complete scene construction, and determines the spatial pose and observation geometry of the relative navigation camera based on the installation matrix and camera parameters. Further, the simulation rendering engine performs layered physical imaging and noise fusion based on an adjustable visible light camera imaging model to generate a continuous image data stream and depth buffer data. Finally, it generates a 3D point cloud and verifies the observation results of the relative navigation algorithm based on the continuous image data stream and depth buffer data, thus forming a closed-loop execution mechanism from dynamics-driven, imaging generation to navigation verification.

[0093] Based on the same inventive concept, embodiments of the present invention also provide a simulation system for relative navigation equipment for intelligent spacecraft, the system module architecture being as follows: Figure 5 As shown. The system includes:

[0094] The dynamics simulation construction module is used to execute the process in step one above, establish a high-precision aerospace dynamics solution module coupled with the simulation engine, and solve the spatial relative positional relationship of major celestial bodies and track the orbital and attitude state parameters of intelligent spacecraft and target intelligent spacecraft in real time under a unified simulation time reference, serving as the underlying data source.

[0095] The space environment simulation construction module is used to execute the process in step two above. It constructs a space environment illumination model based on the spatial relative positions of major celestial bodies and the principle of optical radiation, and constructs a stellar background model based on the star catalog, thus completing the construction of the space environment simulation model.

[0096] The spacecraft system-level simulation construction module is used to execute the process in step three above. It drives the intelligent spacecraft model to perform relative motion mapping based on orbit and attitude state parameters, completes the three-dimensional digital modeling of the intelligent spacecraft based on structural materials, and configures the installation parameters of the relative navigation camera to complete the construction of the spacecraft system-level simulation model.

[0097] The navigation device-level simulation construction module is used to execute the process in step four above. It determines the spatial pose of the relative navigation camera by combining the installation parameters and the attitude of the tracking intelligent spacecraft, and calculates the spatial line-of-sight vector. Based on the spatial line-of-sight vector, it performs layered physical imaging and noise fusion on the target intelligent spacecraft and the star background model to generate a continuous image data stream. Based on the continuous image data stream, it generates a three-dimensional point cloud and uses it to verify the observation results of the relative navigation algorithm, so as to complete the construction of the relative navigation device-level simulation model.

[0098] The specific implementation details of each module have been explained in the method steps and will not be repeated here.

Claims

1. A method for constructing a simulation system for relative navigation equipment of intelligent spacecraft, characterized in that, include: Step 1: Establish a space dynamics solution module coupled with the simulation engine to solve the spatial relative position parameters of the main celestial bodies in real time under a unified simulation time reference, as well as the orbital and attitude state parameters of the tracking intelligent spacecraft and the target intelligent spacecraft, as the underlying data source of the simulation system. Step 2: Based on the spatial relative position parameters of celestial bodies and the principle of optical radiation, construct a space environment illumination model; construct a stellar background model based on the star catalog. The space environment illumination model and the stellar background model together constitute the space environment simulation model. Step 3: Drive the intelligent spacecraft model to perform relative motion based on the orbit and attitude state parameters, construct a three-dimensional model of the spacecraft based on the physical structure and material parameters of the spacecraft, and configure the installation parameters of the relative navigation camera to complete the construction of the spacecraft system-level simulation model; Step four: Based on the installation parameters and attitude state parameters of the relative navigation camera, determine the spatial pose of the relative navigation camera and calculate the spatial line-of-sight vector; configure the physical imaging parameters of the relative navigation camera to establish a visible light camera imaging model; based on the space environment simulation model, spacecraft simulation model, and visible light camera imaging model, perform physical imaging simulation and noise modeling on the target intelligent spacecraft to generate a continuous image data stream; under a unified simulation time reference, synchronously generate and output visible light images, depth maps, point cloud data, and spacecraft orbital attitude state data, establish the correspondence between the observed data and the true state data, and complete the construction of the relative navigation device-level simulation model.

2. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 1, characterized in that, The specific steps for establishing the aerospace dynamics solution module coupled with the simulation engine in Step 1 are as follows: The open-source aerospace dynamics calculation program was modified by removing the graphical interface components and compiling it into a dynamic link library. Based on dynamic link libraries, the aerospace dynamics calculation program is repackaged into a dynamics solution module that can be called by the simulation system; During the simulation system's operation cycle, the dynamics solution module is called periodically according to the logical step size, outputting the spatial relative positions of the main celestial bodies with time alignment, and the orbital and attitude state parameters of the tracking intelligent spacecraft and the target intelligent spacecraft, thereby realizing the spatiotemporal synchronous coupling of the dynamics solution and rendering process.

3. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 1, characterized in that, Step two involves constructing a stellar background model based on a star catalog, specifically including: The star catalog data is preprocessed to extract stellar feature parameters and perform verification, coordinate transformation and format standardization to obtain standardized star catalog data. A celestial sphere spatial index structure is constructed based on the celestial sphere spatial division rules, and a mapping relationship between each star and its corresponding spatial cell is established; the stellar parameters within each spatial cell are pre-calculated and cached to generate a stellar index database. During field-of-view rendering, the spatial index structure is queried based on the field-of-view parameters of the relative navigation camera, spatial cells that intersect with the current field of view are selected, the set of visible stars within the field of view is extracted, and a stellar background model is constructed.

4. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 3, characterized in that, In constructing the celestial spatial index structure, the entire celestial space is regularly divided and the celestial spatial index structure is established; the entire celestial sphere is divided according to the directions of right ascension and declination. The first spatial lattice element, the... The coverage of right ascension and declination of the grid cell is as follows: ； In the formula, Right ascension of a star For the declination of stars, The total number of lattice elements in the right ascension direction. This represents the total number of lattice cells in the declination direction. For cell index, For right ascension index, Index for declination; All stars in the standardized stellar database are loaded into corresponding spatial cells according to their celestial coordinates, forming a stellar index database organized based on celestial spatial location.

5. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 3, characterized in that, The system filters spatial lattices that intersect with the current field of view and extracts the set of visible stars within the field of view. Based on the pose information of the relative navigation camera, it obtains the unit vector of the optical axis direction of the relative navigation camera in real time. and field of view ; Inversely solve the optical axis direction vector into celestial coordinates: ； in: The right ascension is relative to the optical axis of the navigation camera. The declination is relative to the optical axis of the navigation camera. Let be the unit direction vector relative to the optical axis of the navigation camera in the J2000 inertial coordinate system. Using the angular radius, determine the declination and right ascension ranges covered by the current field of view: ； in, Right ascension for field of view coverage, For the declination of the field of view, This is a longitude contraction correction term along the declination direction, used to compensate for the compression effect of the actual coverage angle width of the right ascension lattice cells in high-latitude regions; the calculation of the right ascension range needs to handle the cross-latitude longitude contraction. / Boundary cases; Map the declination and right ascension ranges covered by the field of view to the lattice index: ； in: The lower and upper bounds of right ascension for the field of view coverage The lower and upper boundaries of the declination of the field of view are defined; only the stellar data cached in the cells within the above index range are read to complete the extraction of the set of visible stars in the current field of view, without having to traverse the entire database of stars one by one.

6. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 1, characterized in that, Step three involves constructing simulation models of the tracking intelligent spacecraft and the target intelligent spacecraft, specifically including: The orbital and attitude state parameters are converted into coordinate transformation matrices to dynamically update the three-dimensional spatial pose of the intelligent spacecraft model; A spacecraft surface reflection model is constructed based on the bidirectional reflection distribution function, and material parameters are configured according to the surface material properties of the spacecraft. Establish a direct mapping relationship between material parameters and physical rendering engine material properties. By configuring corresponding material parameters for different structural regions on the spacecraft surface, differentiated rendering of the surface reflection characteristics of different structural components can be achieved. Combining the space environment lighting model, the physical rendering engine performs global lighting calculations on the spacecraft surface based on the configured BRDF model, generating rendered images that conform to real optical properties; The installation parameters of the relative navigation camera are initially configured, including the installation position and pointing direction of the relative navigation camera in the coordinate system of the tracking intelligent spacecraft, and converted into a relative navigation camera installation matrix.

7. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 1, characterized in that, In step four, the specific steps for establishing the visible light camera imaging model and performing physical imaging are as follows: Based on the attitude state parameters of the tracking intelligent spacecraft, construct the attitude transformation matrix of the tracking intelligent spacecraft in the world coordinate system; combine the relative navigation camera installation matrix to calculate the attitude matrix of the relative navigation camera in the world coordinate system; combine the forward unit vector of the relative navigation camera in the local coordinate system to calculate the optical axis direction of the relative navigation camera at the current moment, and obtain the spatial line-of-sight vector of the relative navigation camera. Based on the actual relative navigation camera design specifications, configure physical imaging parameters and establish a visible light camera imaging model; after completing the visible light camera imaging model construction, execute the layered physical imaging process; perform layer compositing on the stellar background image and the target intelligent spacecraft image; and perform noise modeling and degradation processing on the original visible light image.

8. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 7, characterized in that, The 3D point cloud data stream is generated by coupling the depth image from the depth camera in the relative navigation camera with the visible light image from the visible light camera. The specific steps are as follows: Extract the depth values ​​corresponding to each pixel in the depth buffer data generated by the depth camera rendering, and combine the projection parameters and projection inverse matrix of the depth camera to perform spatial recovery of the depth values ​​in order to obtain the actual depth information of the target object in the camera's view space and generate a depth image. The color information of the corresponding pixels in the visible light image is fused with the actual depth information. The graphics processor is used to perform spatial back projection calculation on the effective pixels in the depth image to restore the three-dimensional spatial coordinates of the corresponding pixels in the camera coordinate system. This generates a three-dimensional point cloud data stream that is time-synchronized with the continuous image data stream, so that the relative navigation algorithm can verify the observation results.

9. The method for constructing a simulation system for relative navigation equipment of intelligent spacecraft according to claim 7, characterized in that, To ensure the time consistency between the dynamics calculation process and the graphics rendering process, a unified simulation time management mechanism is established within the simulation system. This mechanism maintains global simulation time variables and updates them incrementally according to logical steps. At the beginning of each simulation cycle, the unified simulation time management mechanism generates a simulation timestamp for the current moment and simultaneously distributes it to the dynamics calculation module, the space environment simulation module, the spacecraft simulation module, and the visible light camera imaging module. Each module performs state updates and image generation processes based on the same timestamp, synchronously outputting visible light images, depth images, 3D point cloud data, and spacecraft orbital attitude state data under a unified simulation time reference, forming a unified time-series multimodal observation data stream.

10. A simulation system for relative navigation equipment of intelligent spacecraft, characterized in that, include: The dynamics simulation building module is used to establish a space dynamics solution module coupled with the simulation engine. Under a unified simulation time base, it can solve the spatial relative positional relationship of major celestial bodies in real time, and track the orbital and attitude state parameters of intelligent spacecraft and target intelligent spacecraft as the underlying data source. The space environment simulation construction module is used to construct a space environment illumination model based on the spatial relative positions of the main celestial bodies and the principle of optical radiation, and to construct a stellar background model based on the star catalog, thereby completing the construction of the space environment simulation model. The spacecraft system-level simulation construction module is used to drive the intelligent spacecraft model to perform relative motion mapping based on the orbit and attitude state parameters, complete the three-dimensional digital modeling of the intelligent spacecraft based on the structural materials, configure the installation parameters of the relative navigation camera, and complete the construction of the spacecraft system-level simulation model. The navigation device-level simulation construction module is used to combine the installation parameters and the attitude state parameters of the tracking intelligent spacecraft to determine the spatial pose of the relative navigation camera and calculate the spatial line-of-sight vector; based on the spatial line-of-sight vector, perform layered physical imaging and noise fusion on the target intelligent spacecraft and the star background model to generate a continuous image data stream; generate a three-dimensional point cloud based on the continuous image data stream, and use it to verify the observation results of the relative navigation algorithm, so as to complete the construction of the relative navigation device-level simulation model.

Images