Spacecrafts face with multiple complex constraints during avoidance maneuvers. There are several problems in the traditional motion planning methods based on numerical optimization when processing corresponding models and constraints, such as the sensitive initial value and long calculation time, which makes it difficult to deal with close range orbital threats in time. To address this problem, a multi constrained avoidance motion planning method based on deep reinforcement learning (DRL) is proposed in this paper. First, the spacecraft six degree of freedom nonlinear dynamical model and related constraints for attitude orbit maneuvers are established. Then, the avoidance motion planning method based on twin delayed deep deterministic policy gradient (TD3) is proposed, and the multi constrained avoidance maneuvering actions can be online generated via the neural networks trained by TD3. Finally, the normative DRL training environment matched with the proposed planning method is constructed to ensure the effective interactions between agents and environments. Simulation results show that the proposed method can rapidly generate avoidance actions in real time when the expected rendezvous time is only in tens of seconds, and the planning period is less than 9 ms, which is much lower than the Gauss pseudo spectral method as a comparison item.

lanetary rover systems need to perform terrain segmentation to identify drivable areas and plan the path, so as to ensure the success of rover detection missions. At present, the task of Mars terrain segmentation is difficult and the computational resources of the rover are limited. This paper proposes a lightweight semantic segmentation network based on DeepLab v3+ network structure. The backbone network is MobileNetV2. The Atrous spatial Pyramid pooling (ASPP) module is optimized by dense connection to further expand the receptive field of the atrous convolution. The coordinate attention (CA) mechanism proposed recently is used to increase the feature extraction ability of our network. CA DeepLab v3+ network is verified by AI4Mars public dataset, which shows that the recall rate of the algorithm can reach 91%, 92%, 89% and 75% in soil, bedrock, sand and large rock, respectively.

A double pulse guidance strategy based on Gaussian perturbation equation is designed for the quick far range rendezvous task after lunar take off. Firstly, the control equation of the orbit correction is derived, and then the double pulse guidance nonlinear equation group is derived by combining the constraint equation of far range rendezvous time. In order to obtain the minimum solution of speed increment, the problem of solving nonlinear equations is transformed into a nonlinear programming problem by designing programming variables, and the optimal solution is solved by sequential quadratic programming algorithm (SQP). In addition, in order to improve the guidance accuracy, the iterative correction method is used to optimize the guidance process. Finally, the correctness of the double pulse guidance strategy based on Gaussian perturbation equation is verified by data simulations, and compared with Lambert double pulse guidance strategy. The simulation results show that the double pulse guidance strategy based on Gaussian perturbation equation can effectively complete the quick far range rendezvous task, and the guidance accuracy and fuel consumption are better than Lambert double pulse guidance strategy.

Facing the problem of high precision image combination and image stabilization adjustment of the primary mirror of the segmented space telescope assembled on orbit, a 6 DOF sub mirror active adjustment mechanism driven by voice coil motor is designed in this paper. Considering the influence of the micro vibration source of the star, a decentralized force control scheme is proposed, which completes high precision control of the mirror's position and attitude. And the image combination and vibration control of the segmented mirror is realized. In this paper, the form of optical sensitivity matrix between aberration and sub mirror pose parameters is derived, and Zernike aberration polynomial of hexagonal aperture is obtained by Gram Schmidt orthogonalization method. Secondly, the whole satellite dynamics model of the segmented space telescope is established. Based on the voice coil motor, a 6 DOF sub mirror active adjustment mechanism is designed, and the position and attitude solutions scheme of the eddy current sensor array are proposed. Then, the control scheme of the segmented space telescope after on orbit assembly is proposed. Aiming at the TTP (tip, tilt, piston) error and micro vibration interference of the sub mirror, the decentralized voice coil motor controller with feed forward compensation and speed feedback is designed via the analytic expression of the position and attitude target quantity of the sub mirror and the expected length of the actuator to adjust the motor output force and the position and attitude of the sub mirror, so as to achieve image combining and vibration control. Finally, the simulation verification and calculation of the imaging quality evaluation index show that the control scheme proposed in this paper can achieve high precision image merging and image stabilization adjustment of the segmented space telescope, with fast adjustment speed, high accuracy and strong vibration attenuation.

Dual Arm space robot can manipulate cooperatively the target with the characters of Heavy Load and high precision, which has the constraint conditions of closed chain and control structure simultaneously. The closed chain dynamics model is provided to achieve the target carry manipulation for dual arm 6 DOF space robot in this paper. Based on the virtual cutoff hinge structure the manipulator arm adopts rigid motion mode, while the cooperative arm adopts compliance motion mode. The joint tracking algorithm and compliance control law are designed respectively. A numerical example of space manipulation verifies the effectiveness of the proposed method.

An impedance control method in which the impedance parameters can be adaptively changed according to the environment is proposed oriented to compliance control of space manipulators in contact with environment when the environmental stiffness and damping change abruptly or gradually. The least square method is used to estimate the environmental damping and stiffness. By analyzing the system characteristics of the space manipulator collided with the environment, a second order system in which the environment and the manipulator are regarded as one is deduced. Based on the response characteristics of the second order system, an adaptive scheme of impedance parameters self adjusting with the environment is proposed. Finally, a numerical simulation is carried out on a planar 2 DOF manipulator. Compared with the traditional impedance control method, the proposed adaptive impedance control has certain advantages in force control accuracy, reducing the maximum contact force and adapting to different contact environments.

Considering the constraint of spacecraft hardware and software resources, a scheduling method for spacecraft limited resources in the process of avoiding abnormal space approaching is proposed for the scenario where spacecraft is abnormally approached by non cooperative space targets. Space situation information under a specific mission is used as input and dynamic spacecraft resource allocation is output. Firstly, the dynamic model of threat avoidance scenario and the limited software and hardware resource model of spacecraft are established to analyze the information flow in the threat avoidance process. On this basis, the “elite reserve” and “inferior elimination” strategies are introduced to design the spacecraft resource scheduling method based on genetic algorithm to accelerate the convergence rate of genetic algorithm. The simulation results show that, compared with the random scheduling strategy, the method proposed in this paper effectively improves the orbit determination convergence rate for non cooperative targets under resource constraints, the mission spacecraft reaches the predetermined position faster, and saves the consumption of incremental velocity.

Constellation imaging scheduling for remote sensing satellites refers to imaging multiple ground target points by scheduling and allocating available satellite observation resources. In order to make the imaging scheduling scheme meet the mission requirements and have the optimal mission efficiency, the constellation imaging scheduling problem is usually formulated as an optimization problem with multiple constraints. In this paper, a satellite digital parallel system is presented, which simulates the operation and control process of satellite constellation in orbit. A digital twin satellite that simulates the functions of GNC subsystem, digital tube subsystem and TT&C subsystem is established in the system, which receives and executes real satellite control commands, and reflects the satellite attitude and orbit control process in the form of telemetry data. Considering the constraints of energy, time window, illumination and meteorological conditions in satellite multi target imaging, the problem of constellation imaging scheduling is modeled as a constrained optimization problem with the target revisit time as the optimization index. The feasible solution set of multi target imaging sequence is obtained through accelerated simulation of satellite digital parallel system, and the constellation imaging scheduling scheme with the highest revisit frequency of multi target points is obtained by solving the optimization problem. The simulation results show that the method proposed in this paper solves the problem of target assignment in constellation imaging scheduling, and the digital twin satellite model adopted can simulate the satellite attitude and orbit control process, thus ensuring that the method is more feasible and timelier in constellation imaging missions.

Ground accompanying flight system of spacecraft is composed of telemetry and telecontrol data interface and high precision digital simulation model. When accompanying flight system is working, telemetry and telecontrol data of the spacecraft are received through data interface in real time and injected into spacecraft model for synchronous simulation, so as to realize the ground digital twin system of the spacecraft in orbit. In this paper, the problem of the design and implementation of telemetry and telecontrol data interface is studied. Firstly, the telemetry and telecontrol data standardization package technology and the bidirectional mapping technology between digital model and data are introduced. Then the design and implementation of data interface are presented, and particularly the data receiving module, data drive and state synchronization module are introduced in detail. Finally, the interface is verified by practical engineering cases. The results show that the telemetry and telecontrol data interface works well and provides effective data support for ground accompanying flight system.

For the problems of great technical difficulty and multi line parallel overlapping of phases in satellite development, a virtual test technology is proposed in this paper. Based on virtual test platform, the onboard software is embedded in the virtual target and runs on a PC together with the dynamic simulation software, parts simulation software and automatic test software. The technology realizes the simulation of real underlying environment, software function and external interfaces. The virtual test can be used in all stages of satellite development, and has multiple modes such as pure software and semi physical, which can not only run with virtual equipment, but also support access to real equipment. Compared with traditional pure hardware tests, virtual test can greatly advance the test time, enrich the test means, reduce test cycles, improve efficiency of development and have lower cost of development, which also has high engineering value.

In the SPC navigation process of asteroid detection, navigation feature selection mainly depends on subjective judgment and causes massive fault tests. In this work, by analyzing the processes of shape model reconstruction, image rendering, template matching and spacecraft localization, we propose four principles of feature selection applying to natural feature tracking during aircraft descent phase, including manufacturability, specificity, light adaptability, and distribution uniformity, etc. The threshold conditions corresponding to the principles are analyzed quantitatively. We validate our principles using real imagery from NASA’s OSIRIS Rex mission to asteroid Bennu, along with the accuracy of local elevations, the correlation performance and the navigation comparison data from NASA NAIF SPICE kernels. The results show that the successful matching rate of rendered features against onboard imagery can reach 96.85%.

The tail problem of star sensor in high dynamic condition reduces the signal to noise ratio and affects the accuracy of the product attitude measurement. This paper designs a method of star sensor image stabilization. The angular velocity of star sensor is obtained by gyro. During the exposure of star sensor, the image stabilization can be realized by using image surface driver to drive the motion compensation of imaging chip. It can reduce or even eliminate the tail, and improve the measurement accuracy of star sensor under dynamic condition. The experimental results show that the accuracy of the star sensor is 57.2% higher than that before the image stabilization compensation. There are obvious advantages over the multi frame superposition algorithm. The Gaussian radius, energy peak and correlation coefficient of star point also improve significantly compared with these indexes before image stabilization. The method can be widely used in star sensors and other photoelectric imaging instruments with high dynamic application scenarios.

In view of the complex signal processing and control process in the whole angle mode control method of hemispherical resonator gyro (HRG), a high precision digital solution has a direct impact on the whole angle mode control and gyro accuracy. The whole angle mode control method of HRG is studied. Taking the non ideal hemispherical resonator as the control object, a whole angle mode control scheme of HRG is proposed. A digital demodulation method based on Kalman filter is proposed to solve the problem of high precision demodulation of signals. By building the whole angle control system model, the simulation results show that the control system can realize fast demodulation and stable control, and verify the feasibility of the digital demodulation scheme and whole angle control scheme.

This article focuses on the difficulty of plasma breakdown during the development of a micro RF ion thruster. Experimental research is conducted on the breakdown characteristics of a low pressure narrow tube inductively coupled plasma source(ICP) using simple present and simple past tenses. The plasma breakdown area in the discharge chamber was studied, and it is found that the breakdown location of the low pressure narrow tube ICPs is not in the antenna area where the electric field is strongest. That is, the breakdown area of the antenna is not in the same area as the plasma maintenance area where the RF ion thruster operates stably, making it difficult to ignite the breakdown using the traditional RF ion thruster structure. Simple present and simple past tenses are used to discuss the results of changes made to the structure of the discharge chamber and the position of the antenna to study the variation of the breakdown conditions. It is found that the above phenomenon can be improved by extending the length of the discharge chamber, increasing the number of antenna turns appropriately, and adjusting the relative position between the antenna and the discharge chamber.