To improve the adaptability of capturing different non-cooperative spacecraft, a novel rigid-flexible coupling mechanism is proposed to adaptively capture the common structural features. And the capture performance is analyzed. Firstly, the capture mechanism consisting of a connecting rod mechanism and elastic grippers is introduced. The common spacecraft structures, such as edges and corners, are clamped by grippers while providing cushioning against impacts during capture. Secondly, the Floating Frame of Reference Formulation (FFRF) is used to establish the rigid-flexible coupling dynamic model of the elastic gripper, considering the nonlinear coupling terms such as tensile pressure and bending. The LuGre friction model is integrated with the Hertz contact force model to refine the contact and collision dynamics model. The complete dynamic model for the capture process is developed by combining the above two models. The simulation results indicate that the proposed mechanism can capture common structural features such as edges and corners. Compared with traditional grippers, this new mechanism could significantly reduce the stable capture contact force, maximum impact force and the number of collisions.

This paper investigates the dynamics modeling and trajectory optimization of a coupled constrained system consisting of a space manipulator and a flexible structure. The objective is to plan the joint-space trajectory that guides the manipulator to a desired configuration, while minimizing vibration and deformation of the flexible structure. The trajectory is simultaneously strictly constrained to the manifold defined by the geometric constraint. Firstly, the augmented Lagrangian approach is adopted to model the coupled constrained dynamics. The space manipulator and the flexible structure are formulated together as a differential-algebraic system with Lagrange multipliers. The multipliers are used as the optimization decision variables to construct the coupled constrained dynamic state equations. Then, the trajectory planning is formulated as an optimal control problem with the geometric constraint and transcribed into a nonlinear optimization problem via a direct collocation method. To address trajectory drift caused by discretization errors, a correction strategy combining the relaxation mechanism and orthogonal projection is proposed. Moreover, a residual-based mesh refinement strategy is introduced. It increases time resolution in segments with high residuals and applies warm-started optimization to improve efficiency. Simulation results show that the proposed method maintains both dynamic consistency and geometric accuracy. It effectively suppresses structural vibration and improves overall feasibility and solver robustness.

Robotic on-orbit assembly calls for controlled manipulation of flexible modules to achieve stable contact with the receptacle structure. This process entails complex dynamics, including base-manipulator coupling, flexible vibrations, and frictional contact. Existing studies often use penalty-based contact laws within rigid-flexible coupled multibody dynamics, which tend to introduce numerical stiffness and restrict time-step sizes in long-horizon simulations. To this end, this paper develops a unified time-stepping framework that integrates a recursive multibody dynamics formulation with an irrotational contact field (ICF) convex contact model, consistently incorporating rigid-flexible coupling and frictional contact into the discrete momentum balance. First, the flexible module is reduced via component mode synthesis (CMS), and a recursive assembly procedure is employed to derive the rigid-flexible coupled multibody dynamics. Second, the framework is extended to frictional contact by embedding contact impulses into the discrete momentum balance, thereby defining the contact-space dynamics. Finally, an ICF-based incremental potential is introduced whose gradient uniquely determines the normal and tangential impulses, recasting frictional contact as an unconstrained convex optimization with improved numerical robustness and discrete-time reversibility. The proposed framework is further evaluated via on-orbit assembly simulations to (i) benchmark free-motion responses against MSC Adams, validating the rigid-flexible coupled dynamics; (ii) assess the proposed contact time-stepping scheme in terms of physical consistency, parametric robustness, and discrete-time reversibility under the same coupled dynamics; and (iii) quantify the sensitivity of impact loads and flexible deformation to key task settings, supporting dynamics assessment and control synthesis.

In conventional detumbling operations, although the debris tumbling angular velocity can be reduced, the normal operation makes it difficult to reduce its angular velocity in the lateral direction, which leads to debris deflection. In view of the characteristics of large disturbance and difficult elimination in the detumbling process of nutation debris, anew detumbling method based on lateral angular velocity peak considering relative pose optimization is proposed. Firstly, the dynamic modeling of the detumbling system with rotatable flexible rod was carried out, motion analysis of the nutation debris conducted. Secondly, the flexible contact process is described, and a contact force model considering friction based on Hertz model, is proposed. Subsequently, the lateral angular velocity of the sailboard debris has a significant influence on the disturbance of the tumbling effect, which may lead to a difficult detumbling, making the debris deflected, and consequently, leading to mission failure. A new method detumbling lateral angular velocity peak considering relative pose optimization is proposed, and a repeated detumbling strategy was designed, which was compared with the conventional strategy based on normal plane. The analysis results indicate that the novel detumbling method based on the peak lateral angular velocity achieves a similar detumbling effect as the conventional method, but with a lower amplitude of detumbling angular velocity. Additionally, the linear displacement and linear velocity of the tumbling target are significantly reduced compared to the conventional strategy. All in all, the new detumbling strategy based on the lateral angular velocity peak can not only achieve the same detumbling efficiency, but also reduce the disturbance to the debris, which has certain advantages in the application scenario of on-orbit service for nutation debris. (c) 2025 COSPAR. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

This paper investigates the action-sequence task planning problem for mobile manipulators in the on-orbit assembly of large-scale space telescopes under complex topological and resource constraints. To address the challenges posed by the tight coupling of topology and resource constraints in space assembly tasks, as well as the poor solvability caused by the exponential growth of the search space with problem scale, we propose a two-stage symbolic planning framework integrating domain knowledge. First, a unified symbolic task-planning representation framework is established, in which Boolean facts and numerical fluents are explicitly modeled and jointly reasoned under a unified semantics, thereby ensuring the logical interpretability and constraint consistency of planning results. Second, a restricted instantiation method based on static relational constraints is proposed, which formulates the satisfaction of static action preconditions as a branch-level constraint satisfaction problem and directly solves valid parameter bindings through relational joins, thus compressing the action-instantiation space at its source. Finally, to address the large number of invalid expansions during search, a feature-triggered reactive candidate action expansion strategy is developed. By selecting highly discriminative dynamic feature predicates through Branch Pruning Information Gain and combining them with inverse indexing for on-demand retrieval of state-relevant actions, the proposed strategy reduces the overhead of invalid applicability checks and improves search efficiency in scenarios with complex constraints. Simulation results demonstrate that, compared with general-purpose methods, the proposed approach can effectively reduce action redundancy, improve planning efficiency and scalability, and maintain stable solution quality across space assembly tasks of different scales. The results further indicate that hierarchically organizing domain knowledge and introducing it into different solving stages is an effective way to improve planning performance for complex space assembly tasks.

In this paper, the compliance control of a flexible-joint manipulator mounted on a flexible base structure is studied for space on-orbit assembly tasks. An adaptive impedance control method is proposed, which integrates environment stiffness estimation and time-scale separation to achieve stable force regulation and compliant interaction. Recognizing the time-scale disparity between the fast motor-side and slow link-side dynamics, the singular perturbation framework is employed to construct a two time-scale dynamic model of the flexible-joint manipulator. Accordingly, a hierarchical control architecture with slow and fast layers is designed. The slow layer generates reference behaviors through a task-space impedance controller driven by interaction force error, and suppresses base vibration through null-space optimization to ensure compliance regulation. The fast layer adopts a torque tracking controller with an adaptive compensation module to improve robustness against model uncertainty. Next, the stability is proven by building a composite Lyapunov function that incorporates slow-layer impedance error and fast-layer tracking error. Simulation results verify that the proposed method achieves effective interaction force regulation and dynamic stability even under uncertain contact stiffness and structural flexibility, demonstrating its applicability in the on-orbit assembly of space structures. (c) 2026 COSPAR. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

With growing on-orbit confrontation complexity, traditional threat assessment for non-cooperative approach observations struggles with subjectivity and latency. We propose a mission-completion-based framework that integrates relative orbital dynamics with optical imaging principles. Relative motion in the orbital plane is parameterized and observation modes are classified using line-of-sight (LOS) coverage and solar illumination. Innovatively, the Structural Similarity Index Measure (SSIM) serves as an objective image-quality metric under varying distance and illumination, aided by a scale-alignment (cropping) procedure to ensure consistent spatial scaling. A Unity-based simulator reproduces dynamic imaging, and threat is quantified by the time-averaged SSIM over an orbital period. Multi-scenario experiments show that the physics-based metric eliminates subjective weighting common to multi-attribute schemes while enabling rapid, interpretable assessment for autonomous optical defense and maneuvering planning.

Powered descent guidance is an enabling technology for extraterrestrial soft landing and reusable rocket recovery. Although the focus of guidance research has tilted toward trajectory optimization recently, feedback tracking control is still necessary to suppress the effects of flight perturbations and uncertainties. In this article, the behavior of the powered descent closed-loop system is considered, and a feedback control synthesis method based on the funnel (finite-time invariant set) is proposed. First, the basic model of the powered descent problem is given, and the state feedback closed-loop system is constructed with uncertainties of thrust and drag. Subsequently, the invariance condition is described as sum-of-squares constraints, and the backward funnel model is established considering closed-loop thrust saturation constraints. Further, to address the multivariate and nonconvex features of the funnel model, a point-by-point solution scheme for funnel control synthesis is formulated integrating the techniques of bilinear search, line search, and heuristic search. The simulation results show that for the considered bounded uncertainties, the states within the funnel inlet will evolve to the specific target set. Compared with linear quadratic regulator (LQR) guidance, the controller obtained by funnel synthesis has a stronger stabilization capability with a certification, which ensures terminal accuracy while enhancing the adaptability to initial state deviations.

As the enabling technology in the on-orbit service and maintenance mission, the space robot assembly technology that realizes the high-precision large space structure assembly mission is the trend of future development. In this paper, for the large space antenna on-orbit assembly task, the design of the space large antenna assembly unit structure and assembly method is completed, and a multi-robot cooperative assembly program is proposed. According to the task requirements, a new type of seven-degree-of-freedom robot with a dual end-effector is designed, and its kinematic modeling and solution are completed. Finally, the motion planning of the robot is realized by using the RT-Connect algorithm, and the reliability of the algorithm is verified in the simulation environment, which demonstrates the ability of the assembly robot to meet the demands of the assembly tasks in the spatially complex environment.

空间环境错综复杂，由航天员出舱完成各类空间任务难度较大。利用空间机器人操控技术，能够扩展航天员的操作能力，有效提高空间作业的自主化和智能化水平。本文介绍了空间作业操控的概念与内涵、国内外空间机器人操控技术应用与发展现状，分析了在轨服务、星表巡视和星表采样等三类场景下空间机器人操控技术的特点。在此基础上，对其中涉及的智能感知与信息融合、任务规划、预测仿真等前沿关键技术进行了总结，并结合未来空间探索的需求，研究提出了空间作业操控技术的智能化发展趋势。

High scientific value areas on celestial bodies such as the Moon and Mars are often located in hazardous terrains. To achieve safe landing exploration, a novel planning method is proposed, which can ensure that the planned trajectory maintains a user-specified distance from obstacles, thus reducing potential collision risk induced by factors such as the body size and model uncertainties. Firstly, the basic model of the trajectory optimization problem and its convexification version is given. The obstacles are modeled as polynomial functions, based on which the “if-then” obstacle avoidance logic explicitly considering the safe distance, is described as a sum of squares constraint. This constraint formulation applies to any obstacle described by a finite number of polynomials, independent of the specific expression of the polynomials (reflecting the shape of the obstacle). Subsequently, the convexification process for the obstacle avoidance constraint is given. Finally, the sequential sum of squares programming problem for the obstacle avoidance trajectory is established, which boils down to a series of semidefinite programming problems. Simulation results show that the closest distance between the planned trajectory and obstacles strictly satisfies the specified distance constraint, and the trajectory could avoid non-convex obstacles. With the promising convergence properties of underlying convex optimization algorithms, advanced autonomous obstacle avoidance guidance schemes are expected to be formed based on the proposed trajectory planning method.

This research focuses on the detumbling planning and control method for the dual-arm space robot post-capturing a non-cooperative tumbling target, which takes the uncertainty of target's inertial parameters and generalized input constraints of robotic system into consideration. Firstly, based on the concept of task compatibility, an efficient detumbling strategy without optimization algorithm is pro-posed, where the target's desired acceleration is in the opposite direction of its velocity with the magnitude determined by scaling factor. Next, a compliance control scheme is designed to track the desired trajectory and the desired contact forces by establishing impedance control for the target and end-effectors. Finally, considering the target's dynamic uncertainties, an adaptive controller is included to compensate the effects of uncertainty and ensure stable detumbling. The simulation results are presented for detumbling a target with inertial parameter uncertainties using a 7 degree-of-freedom dual-arm space robot, which demonstrate the effectiveness of the proposed method.(c) 2023 COSPAR. Published by Elsevier B.V. All rights reserved.

Powered descent guidance is an enabling technology for extraterrestrial soft landing and reusable rocket recovery. Although the focus of guidance research has tilted towards trajectory optimization recently, feedback tracking control is still necessary to suppress the effects of flight perturbations and uncertainties. In this paper, for the Mars landing mission, the behavior of the powered descent closed-loop system is considered and a guidance method based on funnel (finite-time invariant set) scheduling is proposed. Firstly, the basic model of the powered descent problem is given and its state feedback closed-loop system is constructed. Subsequently, considering the thrust uncertainty as well as saturation constraints, the powered descent funnel model is established, and a point-by-point computation method based on Sum of Squares Programming (SOSP) is presented. On this basis, multiple nominal trajectories and their nearby funnels are solved. The guidance scheme based on funnel scheduling is formulated, which utilizes the funnels to provide a good basis for tracking different nominal trajectories, enhancing the adaptability to the initial state dispersions. The simulation results show that for the thrust deviation considered, the states within the funnel inlet can ensure convergence. The similar level of performance on fuel consumption and terminal accuracy of funnel scheduling guidance with convex programming guidance implies the bottleneck in feedback control performance.

To address the problems of low precision and poor real-time performance in the process of part identification and positioning of production line assembly robotic arm, Ghost-SE YOLOv5, an assembly part identification and positioning algorithm integrating lightweight network and attention mechanism is proposed. First, the redundancy of feature map convolution is utilized, which solves the problems of large number of model parameters and floating point operations by using Ghost convolution and Ghost Bottleneck modules. Second, the attention mechanism SE Module is introduced in the backbone network to increase the propensity of feature extraction. Last, the loss function is optimized to speed up the convergence of the model. The results shows that the number of parameters, float operation per second and train time of the proposed algorithm are reduced by 45.98%, 55.99% and 24.07%, respectively. And GPU use was reduced from 7.61G to 6.43G. Furthermore, during the test the precision reached 98.6%, and the recall rate realized 95.3%. The real-time detection performance achieved 97.59 FPS, with an improvement of 34.53%. It can be seen that Ghost-SE YOLOv5 algorithm has better practicality in the part identification and positioning of robotic arm for production line assembly.

An optimization algorithm is presented in this paper for the minimal robust positively invariant (mRPI) set approximations via sums-of-squares (SOS) optimization. The mRPI set is an effective tool for robust analysis of uncertain systems under bounded disturbances. The approximation of the mRPI set is always characterized by a polyhedron computed after finite time iterations. In this paper, an mRPI set is characterized by an ellipsoidal set while bounded parametric uncertainties act on states. The proposed algorithm optimizes the shape matrix of the ellipsoidal set approximation by minimizing the volume of the ellipsoidal set. The algorithm is designed for discrete-time and continuous-time nonlinear systems respectively. The algorithm has the ability to further minimize the mRPI set by optimizing the state-feedback control law. Examples are employed to validate the effectiveness of the proposed algorithms.