Ground verification method for intelligent target identification task of comprehensive power platform

Abstract

The invention provides a ground verification method for an intelligent target identification task of a comprehensive power platform. The ground verification method is implemented based on a relative pose motion simulation platform. According to the method, a high-precision reference is provided through a motion capture positioning system, a service star simulator and a target star model are controlled, and a vivid space relative motion scene is dynamically constructed; a target recognition camera in the service satellite load collects target image information and inputs the target image information to the service satellite comprehensive power platform, and a to-be-verified algorithm is operated; and by comparing an algorithm output result with a reference true value and monitoring the performance of the comprehensive power platform when the algorithm runs, comprehensive quantitative evaluation on the algorithm recognition precision, the recall rate, the robustness and the real-time processing capability on the comprehensive power platform is realized. According to the method, the problems of high-fidelity high-dynamic scene simulation in ground verification of an on-satellite intelligent identification algorithm and accurate performance evaluation of the on-satellite intelligent identification algorithm when the on-satellite intelligent identification algorithm operates on a comprehensive power platform are solved, and key support is provided for reliable in-orbit application of the intelligent target identification algorithm.

Description

Technical Field

[0001] This invention belongs to the field of satellite space mission ground test technology, specifically relating to a ground verification method for intelligent target recognition missions of integrated power platforms. Background Technology

[0002] As aerospace technology develops towards intelligence and autonomy, intelligent satellites with real-time on-orbit perception and decision-making capabilities have become an important trend. Among them, intelligent space target recognition algorithms, as a core technology for achieving autonomous rendezvous, on-orbit servicing, and space situational awareness, directly determine the success or failure of missions through their reliability, real-time performance, and environmental adaptability. Such algorithms need to identify and track space targets in real time and accurately using sensors such as onboard optical cameras under complex space lighting conditions, target characteristics, and relative motion backgrounds.

[0003] To ensure the on-orbit reliability of such intelligent algorithms, it is crucial to conduct thorough and high-fidelity testing and verification during the ground phase. Traditional ground verification methods mainly fall into two categories: one is pure digital simulation, which is flexible but struggles to reproduce the imaging characteristics of real sensors and complex environmental noise; the other is testing based on static or simple motion platforms, which cannot provide the algorithm with dynamic and continuous relative motion scenarios, resulting in insufficient performance evaluation of the algorithm under conditions such as real motion blur, continuous changes in viewpoint, and dynamic transitions in lighting.

[0004] Currently, existing relative pose motion simulation platforms mainly consist of servo-guided rails and robotic arms, and are mostly focused on simulation verification of close-range space manipulation tasks. Servo-guided rail platforms suffer from large size and limited degrees of freedom and range of motion, making it difficult to simulate complex trajectories such as fly-around maneuvers. Robotic arm platforms, on the other hand, are limited by their kinematic singularities and workspace shape, making it difficult to achieve large-scale, omnidirectional, and singularity-free smooth positional motion simulations. Furthermore, both types are primarily designed for rendezvous and docking missions involving cooperative targets, and do not adequately consider the simulation of the optical characteristics and complex attitude changes of non-cooperative targets, or the specific verification needs for intelligent target recognition algorithms. Therefore, the lack of a high-fidelity, high-dynamic ground physical verification environment and systematic method for verifying intelligent target recognition algorithms and their real-time processing performance on satellite integrated electrical platforms is one of the key bottlenecks currently hindering the theoretical design and on-orbit application of intelligent target recognition algorithms. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a ground verification method for intelligent target recognition tasks on integrated power platforms. A high-precision, high-dynamic, and hardware-software co-operational relative pose motion simulation platform is designed as a test environment to verify the intelligent target recognition algorithm and its real-time processing performance on the integrated power platform. This method can directly and effectively evaluate the perception performance and real-time processing efficiency of the algorithm under near-real on-orbit conditions, thereby providing crucial ground test support for the reliable on-orbit application of the algorithm.

[0006] The present invention provides a ground verification method for intelligent target recognition tasks of integrated power platforms, which is implemented based on a relative pose motion simulation platform, the platform including: a service star pose motion simulator, a target star attitude motion simulator and a motion capture and positioning system.

[0007] The service star pose motion simulator includes: a first cooperating arm, a three-axis turntable, a pin chain lifting platform, an omnidirectional AGV trolley, and a first computer.

[0008] Furthermore, the service satellite pose motion simulator described above is a composite robot consisting of the first collaborative arm, a three-axis turntable, a pin chain lifting platform, and an omnidirectional AGV trolley, used to carry and drive the working payload of the service satellite and simulate its motion state in space.

[0009] Furthermore, the omnidirectional AGV controls the service satellite's working load to achieve planar motion; the pin-tooth chain lifting platform controls the service satellite's working load to achieve vertical motion; the first cooperating arm controls the service satellite's working load to achieve high-precision motion trajectory motion; and the three-axis turntable controls the service satellite's working load to achieve posture motion.

[0010] The service satellite payload includes at least a target recognition camera to be verified and an integrated power platform running the algorithm to be verified.

[0011] Furthermore, the target recognition camera is used to acquire target image information, and the integrated circuit platform is used to run the intelligent target recognition algorithm to be verified.

[0012] The target star attitude motion simulator includes: a second cooperating arm, a lifting platform, and a third computer.

[0013] Furthermore, the second collaborative arm is vertically moved by the lifting platform and is only used to adjust the height of the target star's physical model before mission simulation; the second collaborative arm controls the target star's physical model to achieve fixed-point three-axis rotation to generate target features in different attitudes for observation by the target recognition camera on the service star's payload.

[0014] The motion capture positioning system includes: a motion capture camera, a motion capture system truss, and a second computer.

[0015] Furthermore, the motion capture camera acquires the true pose state information of the service satellite pose motion simulator and the target satellite attitude motion simulator in the test space. After analysis and calculation by the second computer, control information is fed back to the service satellite pose motion simulator and the target satellite attitude motion simulator, and a state reference is also provided for the entire test process.

[0016] This invention provides a ground verification method for intelligent target recognition tasks on integrated power platforms, comprising the following steps:

[0017] The working payload is driven by the service satellite pose motion simulator to simulate the 6-DOF motion state of the service satellite in space.

[0018] The target star's physical model is controlled by a target star attitude motion simulator to simulate the target star's 3-DOF attitude state.

[0019] By combining the aforementioned simulators, a 9-DOF relative pose motion simulation is achieved, dynamically constructing a physical scene with complex relative motion for algorithm verification, including: a 3-DOF service star position motion simulation, a 3-DOF service star attitude motion simulation, and a 3-DOF target star attitude motion simulation.

[0020] In dynamic scenarios, the target recognition camera in the service satellite's payload serves as the sole information input for the intelligent target recognition algorithm to be verified, capturing the physical model of the target satellite in real time to obtain raw image information. Simultaneously, the motion capture and positioning system synchronously and independently measures and records the true relative pose values ​​of the service satellite and the target satellite's pose motion simulator.

[0021] Image information acquired by the target recognition camera is input into the integrated power platform, and the spatial target intelligent recognition algorithm to be verified is run for real-time processing. By comparing the algorithm output results with the pose truth value provided by the benchmark measurement system, the performance indicators of the algorithm running on the integrated power platform are analyzed, thereby completing a comprehensive quantitative evaluation of the algorithm's recognition accuracy, recall rate, robustness, and its real-time processing efficiency and resource utilization on the integrated power platform.

[0022] Compared with the prior art, the beneficial effects of the present invention are reflected in:

[0023] (1) This invention proposes a service star relative pose motion simulation platform based on a composite robot, which can realize continuous non-segmented simulation of operation tasks such as close approach, fly-around and follow-up for non-cooperative targets. It provides a high-fidelity test and verification environment for spaceborne intelligent target recognition algorithms that can carry real sensors and simulate complex on-orbit motion conditions. Compared with traditional space mission ground simulation motion simulation platforms, it has stronger mission applicability.

[0024] (2) This invention combines a hardware simulation platform with an algorithm verification process. By introducing the accurate benchmark truth value provided by the motion capture system and real-time monitoring of the integrated power platform, a high-precision quantitative evaluation system is established. This system can efficiently and accurately test the accuracy, recall, and robustness of the algorithm to be verified, as well as its real-time processing performance on the integrated power platform, significantly improving the sufficiency and reliability of ground testing. Attached Figure Description

[0025] Figure 1 This is a flowchart of the ground verification method for intelligent target recognition tasks of integrated power platforms proposed in this invention.

[0026] Figure 2 This is a schematic diagram of the relative pose motion simulation platform on which the ground verification method for intelligent target recognition tasks of integrated electric platforms proposed in this invention is based.

[0027] Figure 3 This is a schematic diagram of the coordinate system of the first cooperative arm base and the coordinate system of the service satellite working payload of the present invention's service satellite pose motion simulator.

[0028] Figure 4 This is a schematic diagram of the target star physical model body coordinate system of the target star attitude motion simulator of the present invention.

[0029] In the figure, 1-first collaborative arm, 2-three-axis turntable, 3-motion capture camera, 4-target satellite physical model, 5-second collaborative arm, 6-motion capture system truss, 7-lifting platform, 8-service satellite working payload, 9-omnidirectional AGV trolley, 10-pin chain lifting platform, 11-first computer, 12-second computer, 13-third computer. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention are described clearly and completely below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0031] This invention proposes a ground verification method for intelligent target recognition tasks on an integrated power platform. This method is implemented based on a relative pose motion simulation platform, which includes a service satellite pose motion simulator and a target satellite attitude motion simulator. The principle of this method is as follows: a motion capture and positioning system is used to measure the pose states of the service satellite and the target satellite in real time and independently, providing navigation and positioning information for the service satellite pose motion simulator, which is then used as a benchmark for evaluating algorithm performance. The service satellite pose motion simulator drives its payload to achieve six-degree-of-freedom motion simulation of the service satellite. The target satellite attitude motion simulator controls the physical model of the target satellite to achieve three-degree-of-freedom attitude simulation of the target satellite. This constructs a highly dynamic test and verification scenario. In the dynamic scenario, the target recognition camera on the service satellite's payload acquires target images, which are input into the integrated power platform to run the algorithm to be verified. By comparing the algorithm output with the aforementioned benchmark and monitoring the performance of the integrated power platform during algorithm execution, a comprehensive evaluation of the algorithm's recognition capability and its processing capability on the integrated power platform is completed.

[0032] Combination Figures 1-4 The ground verification method for intelligent target recognition tasks of integrated power platforms described in this invention comprises the following steps:

[0033] Step 1: Configure the service star pose motion simulator, the target star attitude motion simulator and the motion capture and positioning system, build a high-fidelity test environment and calibrate the reference coordinate system;

[0034] The service satellite pose motion simulator includes: a first cooperating arm 1, a three-axis turntable 2, an omnidirectional AGV trolley 9, a pin chain lifting platform 10, and a second computer 12. The service satellite payload 8 is mounted on the three-axis turntable 2, and it includes at least a target recognition camera and a satellite integrated power platform running the intelligent target recognition algorithm to be verified.

[0035] The target star attitude motion simulator includes: a lifting platform 7, a second cooperating arm 5, and a third computer 13.

[0036] The motion capture positioning system includes: a motion capture camera 3, a motion capture system truss 6, and a second computer 12.

[0037] Combination Figure 2 A toothed chain lifting platform 10 is fixed to an omnidirectional AGV trolley 9; a first cooperating arm 1 is fixed to the toothed chain lifting platform 10; a three-axis turntable 2 is fixed to the working end of the first cooperating arm 1; a motion capture camera 3 is mounted on a motion capture system truss 6, which is installed around the experimental area. During the test and verification process, a second computer 12 receives data from the motion capture camera 3 and calculates the true values ​​of the pose states of the service satellite and the target satellite, then transmits this data to the first computer 11 to calculate the control commands for the service satellite simulator.

[0038] Specifically, the omnidirectional AGV trolley 9 simulates the large-scale motion of the service satellite's working payload 8 in the XY horizontal plane; the pin-tooth chain lifting platform 10 simulates the large-scale motion of the service satellite's working payload 8 in the vertical direction of the Z-axis; the first cooperative arm 1 simulates the small-scale high-precision motion of the service satellite; and the three-axis turntable 2 simulates the attitude motion of the service satellite. These motions together generate a dynamic relative motion scenario for testing the algorithm.

[0039] Specifically, an optimal working area is set for the first collaborative arm 1, ensuring that it can move freely, respond quickly, and has no singularities within this area. The omnidirectional AGV trolley 9 and the toothed chain lifting platform 10 move along the XYZ axes, ensuring that the optimal working area always covers the short-term motion trajectory of the service satellite's working payload 8. A simulation test reference coordinate system is defined. First Collaborating Arm 1 Base Coordinate System With the service satellite's working payload in body coordinate system 8 Two-dimensional images of the service satellite motion simulator are acquired by multiple motion capture cameras 3. Based on the principle of multi-view computer vision, the absolute state information of the base coordinate system of the first cooperating arm 1 and the body coordinate system of the service satellite working payload 8 in the simulation test reference coordinate system is obtained by the second computer 12. Through coordinate system transformation, the relative state of the body coordinate system of the service satellite working payload 8 and the base coordinate system of the first cooperating arm 1 is represented in the base coordinate system of the first cooperating arm 1 and provided to the first computer of the first cooperating arm 1, ultimately realizing high-precision control of the motion trajectory of the service satellite working payload 8. At the same time, the above pose information is used as a reference true value for comparison with the output results of the intelligent recognition algorithm to evaluate the algorithm performance.

[0040] Combination Figure 2 Install the second collaborative arm 5 on the lifting platform; install the target star physical model 4 on the working end of the second collaborative arm 5.

[0041] Specifically, the target star physical model 4 body coordinate system is defined. And fix it in the reference coordinate system The target position in the simulation; the coordinate system of the target star physical model 4 deflects relative to the reference coordinate system as the mission simulation time progresses. The second cooperative arm 5 is controlled by the third computer to realize the deflection of the coordinate system of the target star physical model 4 relative to the reference coordinate system and the attitude movement of the target star physical model 4 around its body coordinate system in pitch, yaw and roll directions, providing a dynamic observation target with different perspectives and features for the recognition algorithm to be verified.

[0042] Step 2: Based on the real-time calculation of the service satellite and target satellite state information by the dynamic simulation computer, drive the service satellite and target satellite pose motion simulator to simulate the on-orbit relative motion state in the ground experimental environment, and dynamically construct the mission test scenario.

[0043] Adjust the target star attitude motion simulator and move the second cooperating arm 5 to a suitable spatial height via the lifting platform 7;

[0044] The position, attitude, and reference coordinate system of the target star's physical model 4 were recalibrated using a motion capture and positioning system. And make the service satellite pose motion simulator control the service satellite working load 8 to move to the simulation start state;

[0045] The dynamic simulation computer calculates and generates the state information of the service satellite and the target satellite under simulated on-orbit conditions. The first computer 11 and the third computer 13 drive the service satellite pose motion simulator and the target satellite attitude motion simulator respectively to simulate the dynamic motion scenario in the experimental space, thus constructing a dynamic test scenario for algorithm verification.

[0046] Specifically, the orbital system of the target star (VVLH system) is used as the simulation reference coordinate system.

[0047] Specifically, Yaw angle The pitch angle, This is the roll angle.

[0048] Calculate the simulator attitude transformation matrix according to the rotation sequence in 3-1-2. for:

[0049] , This indicates the attitude rotation caused by the yaw angle. This indicates the attitude rotation caused by the pitch angle. This indicates the attitude rotation caused by the roll angle.

[0050] ;

[0051] For the service satellite's working payload body coordinate system In the first cooperative arm base coordinate system The coordinates of the origin in the equation are: With attitude transformation matrix T represents transpose;

[0052] Therefore, the coordinate system of the service satellite's working payload is used. To the first cooperative arm base coordinate system The conversion relationship is as follows:

[0053] ;

[0054] For the first cooperative arm base coordinate system In the simulation test reference coordinate system The coordinates of the origin in the equation are: With attitude transformation matrix .

[0055] Therefore, based on the coordinate system of the first collaborating arm base To the simulation test reference coordinate system The conversion relationship is as follows:

[0056] ;

[0057] Calculate the angle control values ​​of the service satellite motion simulator according to the rotation sequence of 3-1-2:

[0058] ;

[0059] Among them, the attitude transformation matrix .

[0060] Calculate the displacement control quantity of the service satellite motion simulator:

[0061] ;

[0062] For the simulation test reference coordinate system Second cooperative arm base coordinate system The coordinates of the origin in the equation are: With attitude transformation matrix .

[0063] Therefore, based on the coordinate system of the second cooperative arm base To the simulation test reference coordinate system The conversion relationship is as follows:

[0064] ;

[0065] For the second cooperative arm base coordinate system In the target star physical model body coordinate system The coordinates of the origin in the equation are: With attitude transformation matrix .

[0066] Therefore, based on the target star's physical model body coordinate system To the second cooperative arm base coordinate system The conversion relationship is as follows:

[0067] ;

[0068] Therefore, based on the target star's physical model body coordinate system To the simulation test reference coordinate system The conversion relationship is as follows:

[0069] ;

[0070] Calculate the angle control values ​​of the target star motion simulator according to the rotation sequence of 3-1-2:

[0071] ;

[0072] Among them, the attitude transformation matrix .

[0073] Calculate the displacement control parameters of the target star motion simulator:

[0074] ;

[0075] Therefore, the service satellite's working payload body coordinate system To the target star physical model body coordinate system The conversion relationship is as follows:

[0076] ;

[0077] Then, the relative displacement state in the control of the simulation system Transformed into the overall displacement state of the service satellite for:

[0078] ;

[0079] Relative attitude state Transformed into the overall attitude state of the service satellite for:

[0080] .

[0081] Based on this, the relative displacement and relative attitude states of the simulation system can be converted into the comprehensive displacement and comprehensive attitude control quantities required by the service satellite simulator and the target satellite simulator, thereby realizing the accurate simulation and control of the entire dynamic test and verification scenario.

[0082] Step 3: The service satellite payload carried by the service satellite pose motion simulator includes at least a target recognition camera and satellite integrated circuit, acquires target image information, and runs the intelligent target recognition algorithm to be verified.

[0083] During dynamic scenario operation, the target recognition camera on the service satellite's payload 8 serves as the sole sensor for the algorithm under test, capturing real-time images of the target satellite's physical model 4 to obtain raw image sequences. The image data acquired by the target recognition camera is then input into the integrated power platform within the service satellite's payload 8 in real-time. This platform runs the space target intelligent recognition algorithm to be verified under simulated real-world spaceborne resource constraints. Key performance indicators during the algorithm's processing are monitored and recorded simultaneously, including single-frame processing latency, frame rate, and CPU and memory usage.

[0084] Step 4: Using the motion capture and positioning system in the platform, synchronously and independently measure and record the precise relative pose information of the service satellite and the target satellite in the task verification scenario, and use this as the benchmark true value for evaluating the algorithm performance.

[0085] Throughout the testing process, the motion capture and positioning system operated independently of the algorithm flow, synchronously acquiring the spatial positions of marker points installed on the service satellite's payload and marker points on the target satellite's physical model at a frequency higher than the algorithm's processing frequency. Utilizing multi-view vision principles, the second computer 12 calculated in real-time the high-precision position and attitude information of both points in the simulation test's reference coordinate system, and further calculated their precise relative pose. This information was strictly synchronized with the target recognition camera's image acquisition, forming an objective baseline value corresponding to each frame of the algorithm's input image, which was recorded and stored for subsequent performance evaluation.

[0086] Step 5: Compare and analyze the output information of the intelligent target recognition algorithm to be verified with the synchronous benchmark truth value recorded in Step 4, evaluate the recognition accuracy, recall rate and robustness of the tested algorithm in real time, and determine whether the current task has ended. If it has ended, proceed to Step 6; otherwise, return to Step 2.

[0087] The results output by the intelligent recognition algorithm in step 3, including target recognition confidence, category, and coarse pose, are compared and analyzed with the synchronous reference ground truth provided by the motion capture and positioning system in step 3. The recognition accuracy, recall and robustness of the algorithm under different dynamic conditions are evaluated in real time and it is determined whether the current task has ended. If it has ended, proceed to step 6; if it has not ended, return to step 2 and re-execute.

[0088] Step 6: Analyze the key performance indicators of the integrated power platform during the operation of the intelligent target recognition algorithm, and evaluate the real-time processing efficiency and resource utilization of the tested algorithm under the constraints of the real integrated power platform.

[0089] After the mission is completed, the key performance indicators of the integrated power platform during the mission processing are analyzed, including single-frame processing latency, frame rate, CPU and memory usage. Combined with the algorithm performance evaluation results from step 5, a comprehensive quantitative evaluation report is finally generated on the algorithm's recognition performance and its real-time processing efficiency and resource utilization on the spaceborne integrated power platform.

Claims

1. A ground verification method for intelligent target recognition tasks on integrated power platforms, implemented based on a relative pose motion simulation platform, characterized in that, Includes the following steps: Step 1: Configure the service star pose motion simulator, the target star attitude motion simulator and the motion capture and positioning system, build a high-fidelity test environment and calibrate the reference coordinate system; Step 2: Based on the real-time calculation of the service satellite and target satellite state information by the dynamic simulation computer, drive the service satellite and target satellite pose motion simulator to simulate the on-orbit relative motion state in the ground experimental environment, and dynamically construct the mission test scenario. Step 3: The service satellite payload (8) carried by the service satellite pose motion simulator includes at least a target recognition camera and satellite integrated circuit, collects target image information, and runs the intelligent target recognition algorithm to be verified. Step 4: Using the motion capture and positioning system in the platform, synchronously and independently measure and record the precise relative pose information of the service satellite and target satellite pose motion simulator in the task verification scenario, and use this as the benchmark true value for evaluating the performance of the intelligent target recognition algorithm. Step 5: Compare and analyze the output information of the intelligent target recognition algorithm to be verified with the synchronous benchmark truth value recorded in Step 4, evaluate the recognition accuracy, recall rate and robustness of the algorithm in real time, and determine whether the current task has ended. If it has ended, proceed to Step 6; otherwise, return to Step 2. Step 6: Analyze the key performance indicators of the integrated power platform during the operation of the intelligent target recognition algorithm, and evaluate the real-time processing efficiency and resource utilization of the tested algorithm under the constraints of the real integrated power platform.

2. The ground verification method for intelligent target recognition tasks of integrated power platforms according to claim 1, characterized in that, The service satellite pose motion simulator includes: a first collaborative arm (1), a three-axis turntable (2), a pin chain lifting platform (10), an omnidirectional AGV trolley (9), and a first computer (11); the first collaborative arm (1), the three-axis turntable (2), the pin chain lifting platform (10), and the omnidirectional AGV trolley (9) constitute a composite robot, which is used to carry and drive the service satellite's working payload (8) and simulate its motion state in space.

3. The ground verification method for intelligent target recognition tasks of integrated power platforms according to claim 1, characterized in that, The target star attitude motion simulator includes: a second cooperative arm (5), a lifting platform (7), and a third computer (13); the lifting platform (7) controls the vertical movement of the second cooperative arm (5), and is only used to adjust the height of the target star physical model before the mission simulation; the second cooperative arm (5) controls the target star physical model to achieve fixed-point three-axis rotation to generate target features under different attitudes for observation by the target recognition camera on the service star's working payload (8).

4. The ground verification method for intelligent target recognition tasks of integrated power platforms according to claim 1, characterized in that, The service satellite payload (8) includes at least a target recognition camera and an integrated power platform for running the intelligent target recognition algorithm to be verified; the target recognition camera is used to acquire target image information, and the integrated power platform is used to run the intelligent target recognition algorithm to be verified.

5. The ground verification method for intelligent target recognition tasks of integrated power platforms according to claim 1, characterized in that, The motion capture positioning system includes: a motion capture camera (3), a motion capture system truss (6), and a second computer (12). The motion capture camera (3) acquires the true value information of the pose state of the service star pose motion simulator and the target star attitude motion simulator in the test space. After analysis and calculation by the second computer (12), control information is fed back to the service star pose motion simulator and the target star attitude motion simulator. At the same time, it provides a state reference for the entire test process.

6. The ground verification method for intelligent target recognition tasks of integrated power platforms according to claim 5, characterized in that, When constructing the test scenario, the three-degree-of-freedom positional motion of the service satellite is simulated by the composite robot, and the three-degree-of-freedom attitude motion of the service satellite is simulated by the three-axis turntable (2).

7. The ground verification method for intelligent target recognition tasks of integrated power platforms according to claim 1, characterized in that, When constructing the test scenario, the second collaborative arm (5) controls the target star physical model (4) to achieve fixed-point three-axis rotation in order to simulate the three-degree-of-freedom attitude motion of the target star.

Images