Reliable depth perception is critical for robotic manipulation, especially for non-Lambertian objects such as transparent or highly specular surfaces, where raw depth measurements are often corrupted or missing. These failures frequently propagate to motion planning, resulting in invalid grasp poses and execution errors. We propose AISPO, a depth completion framework that improves depth reliability for manipulation in challenging sensing conditions. AISPO combines multi-scale RGB-D feature fusion with an affine-invariant shape prior to enforce geometric consistency and mitigate catastrophic depth failures. Unlike methods that focus primarily on average depth accuracy, our approach emphasizes physical plausibility and structural integrity of the predicted depth maps. Extensive benchmark evaluations demonstrate competitive performance and strong generalization to unseen objects and novel scenes. Real-world grasping experiments further show that enhanced depth reliability significantly improves manipulation success rates, particularly for transparent objects where many existing methods fail to produce physically usable depth estimates.

Reliable fall recovery, which commonly aims at attaining a nominal upright posture, is essential for the autonomous operation of humanoid robots in unstructured field environments. Although existing posture-centered methods can synthesize coordinated whole-body recovery motions from diverse fallen configurations, they may result in a dynamically fragile support state, leading to secondary loss of balance or unstable resumption of commanded locomotion, particularly under terrain-dependent contact conditions. We propose to learn a unified humanoid policy from fall recovery to locomotion (UniReLo) across heterogeneous field terrains. UniReLo leverages continuously gated multi-scale motion priors to modulate frame-, sequence-, and gait-level adversarial supervision according to recovery progress, preserving the distinct temporal structures of recovery and locomotion without requiring fixed-threshold switching. In addition, terrain-conditioned recovery guidance evaluates the evolving support state using a terrain-relative support representation and support-feasibility assessment. Simulation and outdoor real-world experiments demonstrate that UniReLo can deliver stable and continuous recovery-to-locomotion behaviors for humanoids across diverse field terrains. The supplementary video is available at https://vsislab.github.io/UniReLo/.

Solid-state refrigeration technology has extensive applications across electronics, optics, aviation, and aerospace, due to its environmental compatibility, precise temperature regulation, and compact structuring. However, in complex and volatile environments, variations in thermoelectric characteristics lead to fluctuations in the temperature difference performance of thermoelectric coolers (TECs), making the prediction of extreme temperature more challenging. To address the challenge of achieving precise control over the extreme temperature of TECs across diverse environments, an optimized control method is proposed. This method combines a TEC performance predicting model with a fuzzy logic control strategy. The genetic algorithm (GA) optimized back propagation neural network (BPNN) prediction model is developed to provide the initial target for extreme temperature control. To improve prediction accuracy, a novel expression of TEC refrigeration performance is proposed. A fuzzy control strategy is introduced to identify a lower and more precise temperature target across diverse environments. Based on the temperature descent gradient, fuzzy logic rules are formulated to optimize extreme temperature regulation, specifically addressing TEC's inherent strong nonlinearity. Finally, experiment results confirm the effectiveness: compared to the sole GA-BPNN model, the proposed method further reduces the extreme cooling temperature by 4.3 degrees C and maximizes the coefficient of performance (COP) with a 12.57% improvement. Additionally, under the optimized BP-PID (OBP-PID) strategy with manually specified target from the proposed method, the extreme temperature is achieved only under a 20 degrees C ambient temperature. Meanwhile, the COP of the proposed method exhibits a 22.56% enhancement relative to the OBP-PID strategy. This proposed control method enables the full utilization of TEC performance, thereby providing critical impetus for the practical advancement of solid-state refrigeration technology.

This paper proposes a control method for space manipulators in zero-gravity to grasp non-cooperative targets. A task-specific observation space, joint-angle control action mapping, and hierarchical reward function were devised to guide the policy through approach, alignment, and grasp stages. Experiments were carried out with the Franka Emika Panda arm in NVIDIA’s Isaac Lab simulation platform. Results show the policy can accurately track target trajectories, reach millimeter-level precision, and converge quickly during training, demonstrating robustness and efficiency. The presented approach provides a feasible solution for autonomous space manipulator operations, with potential applications in space station maintenance, on-orbit servicing, and space debris removal.

Observable conditions are given for tracking and observing non-cooperative targets under space-based single-star angular-only tracking. To solve the degradation of the estimation accuracy of unscented Kalman filtering (UKF) when the observable degree is low, this paper adopts an improved unscented Kalman filtering (IUKF) algorithm based on observable degree. With the introduction of the energy scaling parameter, the filter gain covariance matrix can be adjusted online according to the size of the energy scaling to modify the weights of state prediction and system observation in real time. Mathematical simulations show that the application of this method to low-orbit observation satellites can improve the estimation accuracy and shorten the convergence time, which can be better applied to near-Earth satellite observation missions with short rendezvous times.

3D point cloud maps are widely used in robotic tasks like localization and planning. However, dynamic objects, such as cars and pedestrians, can introduce ghost artifacts during the map generation process, leading to reduced map quality and hindering normal robot navigation. Online dynamic object removal methods are restricted to utilize only local scope information and have limited performance. To address this challenge, we propose DORF (Dynamic Object Removal Framework), a novel coarse-to-fine offline framework that exploits global 4D spatial-temporal LiDAR information to achieve clean static point cloud map generation, which reaches the state-of-the-art performance among existing offline methods. DORF first conservatively preserves the definite static points leveraging the Receding Horizon Sampling (RHS) mechanism proposed by us. Then DORF gradually recovers more ambiguous static points, guided by the inherent characteristic of dynamic objects in urban environments which necessitates their interaction with the ground. We validate the effectiveness and robustness of DORF across various types of highly dynamic datasets.

Thermoelectric cooler (TEC) is widely used for temperature control in optoelectronics, machinery, biomedicine, and other fields. However, the unstable temperature control over wide ranges and different gradients restricts the efficiency of TEC. To highlight the neural network's effectiveness in controlling TEC systems, the optimized BPPID (OBPPID) strategy is proposed. In the OBPPID strategy, the factors s and G are decided through formula derivation and the particle swarm optimization (PSO) algorithm, respectively contributing to the backward adjustment and forward propagation of BPPID. The OBPPID solves three issues that the BPPID faces: gradient vanishing of network weights, over-reliance of initial network weights and range constraint of control parameters. The OBPPID and BPPID have been simulated 1000 times to evaluate control performance. The results show that, with random initial network weights, the OBPPID strategy reduces the cost variability of the BPPID strategy from 92 to 17, achieving an 81.52% reduction. Physical tests have confirmed that in the temperature control over wide range and the temperature gradients of 5 degrees C, degrees C, 20 degrees C, degrees C, and 80 degrees C, degrees C, the OBPPID strategy can adaptively adjust the control parameters and provide higher efficiency than PID, BPPID and fuzzy PID strategies for TEC system.

The presented study focuses on the experimental system identification for the longitudinal aircraft model of a micro fixed-wing unmanned airplane in ground effect flights and the longitudinal dynamic stability analysis. The primary contribution is modifying the traditional longitudinal aerodynamic models (i.e. lift coefficient CL, drag coefficient CD and pitching moment coefficient Cm) to incorporate the influence of altitude variation in the ground effect and utilize experimental flight data to identify the model parameters. In this study, we incorporate the relative altitude as an additional state variable. This extends the ordinary Taylor expansion of the longitudinal aerodynamic coefficient model to include the "height derivatives" (i.e. CLh, CDh and Cmh [4]). Agenetic optimization algorithm is applied to the flight data and executes the estimation of the model derivatives. After filling the aircraft state space model, the primary observation was that the Phugoid mode becomes unstable, and the Short Period Mode turns out to be less stable in the ground effect flight. Meanwhile, an unconventional convergent and non-oscillatory mode shows the aircraft's tendency to maintain its nominal flight speed and altitude. It should be noted that different aircraft configurations and designs could lead to different values of height derivatives, which will consequently lead to different ground effect dynamic behaviour. For the illustration, a root locus analysis with respect to each height derivative is performed. The single impact of each height derivative is observed so that it provides insights into the aircraft design considerations.

Large-scale 3D mapping is an important task for robotics and autonomous driving. However, mobile robots and autonomous vehicles with limited hardware resources may face issues with large memory consumption. It is challenging to achieve a balance between mapping quality and memory consumption. To address this issue, we propose a new compact implicit neural map representation - the Tri-Pyramid that can infer the Truncated Neural Distance Field (TNDF) given an arbitrary 3D position. Additionally, we introduce a TNDF label rectification method considering both the direction of ground normals and closest surface points to enhance the precision of supervision signals for training with a set of effective loss functions. Experiments on public datasets demonstrated that our method reaches comparable or superior performance for dense mapping while significantly reducing memory consumption compared to previous LiDAR mapping approaches. Furthermore, our study confirms the scalability and adaptability of our approach from room-scale to city-scale scenes. Moreover, we explore the potential of directly leveraging the implicit neural map representation for localization tasks by solving an optimization problem. The experiments showcase the accurate localization capabilities of our method in various scenarios.

A large number of sensors carried by spacecraft are facing the risk of irradiation, which may lead to the failure of sensors and affect the mission of the spacecraft. Therefore, it is significant to make the sensor work safely by attitude maneuvers. This article proposes a guidance algorithm for the attitude reorientation of rigid spacecraft, which aims to reduce the conservativeness of the traditional potential function and to solve the problem of the goal being non reachable with obstacles nearby (GNRON). While the rationality of the form design of the potential function is analyzed from the view of numerical calculation, a modified inverse-proportional potential function is designed based on it. Additionally, the anti-unwinding attitude error function is established to avoid the unwinding phenomenon of quaternions. In this paper, a set of weight assignment schemes for the potential function is proposed to obtain the maneuver path close to the edge of the attitude-forbidden zone, which reduces the conservativeness and energy consumption compared with the traditional method. Finally, the scheme can also solve the GNRON phenomenon. In the fourth section, the form design of the potential function is studied by numerical simulation, and the performance of the proposed algorithm is shown by comparative experiments.

Detecting moving events produced by moving objects is a crucial task in the realms of autonomous driving and mobile robots. Moving objects have the potential to create ghost artifacts in mapped environments and pose risks to autonomous navigation. LiDAR serves as a vital sensor for autonomous systems due to its ability to provide dense and precise range measurements. However, existing LiDAR datasets often lack sufficient discussion on the motion labeling of moving objects, containing only a limited representation of moving entities within a single scene. Furthermore, the methodologies for Moving Event Detection (MED) on LiDAR sensors have not been comprehensively explored or evaluated. To address these gaps, this study focuses on constructing a diverse LiDAR moving event dataset encompassing multiple scenes with a high density of moving objects. A thorough review of current MED techniques is conducted, followed by the establishment of a performance benchmark based on evaluating these methods using our dataset. Additionally, part sequences of the dataset are utilized to host an online MED competition, aimed at fostering collaboration within the research community and advancing related studies.

The non-overshooting conditions measured by the state/output 2-norm of discrete-time switched linear systems(DTSLS) are investigated in this paper. A sufficient and necessary condition for state non-overshooting of discrete-time linear time invariant(LTI) systems is firstly derived, and the non-overshooting region, which is a pair of zero-symmetrical cones, is defined and analyzed. Adopting the idea of dividing the state space into distinct non-overshooting regions, some state non-overshooting conditions for DTSLS are addressed. The output non-overshooting problem, which is more complex, is solved in a similar way, and several output non-overshooting conditions are established. Finally, an illustrative example is given to verify the proposed conditions.

The number of space targets in the near-Earth orbit has greatly increased, and space-based radar has the advantage of high resolution and high accuracy tracking to enhance the tracking efficiency of space extended targets (SET). We propose a gamma Gaussian inverse Wishart based on a matched linearization-Poisson multi-Bernoulli mixture (GGIWML-PMBM) filter to estimate the motion state and shape size of the SET. For the random matrix that can only describe the distribution under which the measurements are linear, the measurements of the spatial target are linearized using matched linearization, and the extended information is preserved in the second-order central moments. The transfer density and likelihood function of GGIW are nonlinear for Poisson measurements with nonlinear Gaussian spatial distributions, and the single-target densities and normalizing constants of the PMBM filter are not closed form. The prediction and update of PMBM include Gaussian-weighted integral calculation, for which different nonlinear approximation methods are used to calculate the weighted integral and derive the closed form of GGIWML-PMBM. Finally, the simulation scenarios of low-orbit single-radar sensor tracking near SET and group SET are established, and the results show that the tracking accuracy can reach 2.6 m for near SET and 6.6 m in group SET.

In this paper, a multi-prior mixture (MPM) distribution and arithmetic average (AA) fusion-based Student’s t filter (MPMAASTF) is proposed for the problem of non-stationary heavy-tailed noises filtering with only rough prior information available. Firstly, a MPM distribution is proposed to model the joint probability density function (PDF) for one-step prediction and measure likelihood, then the Kullback-Leibler Divergence (KLD) minimization technique and Jensen’s inequality are used to derive approximate posterior PDFs for states and parameters. A finite number of Student’s t PDF fusion algorithm based on AA fusion is developed to guarantee the closure of the filtering posterior PDF, and finally existing high precision moment matching technique is embedded. The proposed filter improves the filtering accuracy of non-stationary heavy-tailed noises by adaptive learning of the degrees of freedom (dof), the one-step prediction error scale matrix and the measurement noise scale matrix. Simulations validate the effectiveness and superiority of the proposed method.

In Kalman filter-based spacecraft autonomous celestial navigation, Inertial/Celestial integrated navigation system and asynchronous multi-sensor data fusion, complex environments and sensor alignment errors are likely to lead to inaccurate statistical priors for the noise covariance matrices and the non-zero measurement noise mean vector (MNMV). To address this issue, this paper firstly proposes a MultiNormal-Inverse Wishart (MNIW) mixture distribution modelling the joint probability density function (PDF) for one-step prediction and measure likelihood, then the MNIW mixture distribution is decomposed into a Gaussian hierarchy, and finally the posterior estimates of the state and variables are obtained using variational Bayesian technique. In this paper, a Multi-Normal-Inverse Wishart mixture distributionbased variational Bayesian extended Kalman filter (VB-EKF) is proposed, in which a first-order Taylor expansion is used to address the non-linear problem. The proposed filter can be used to address nonlinear filtering problem with inaccurate noise covariance matrices and measurement bias. Simulations of spacecraft autonomous celestial navigation validate the effectiveness and superiority of the proposed filter.

In the navigation and positioning process of global navigation satellite system (GNSS) in complex environment, due to occlusion, multipath effect and other faults, the navigation and positioning efficiency of GNSS will be reduced. Aiming at the problem that GNSS system is insensitive to fault tolerance and robust, a batch covariance estimation (BCE) navigation algorithm based on factor graph optimization is proposed. In this paper, firstly, by constructing the factor graph model of GNSS observation data, the graph optimization algorithm is applied to GNSS navigation to improve the navigation accuracy. Then, a BCE algorithm based on factor graph is proposed for GNSS fault, and the operation flow is designed to improve the navigation robustness. Finally, GNSS measurement data with different performance are used for simulation, and noise is added to the data to evaluate the method proposed in this paper. The results show that the BCE algorithm based on factor graph is better than the traditional methods in positioning accuracy and fault tolerance, especially in the case of low-performance measurement, which can significantly reduce the positioning error, and has the ability of fault tolerance when there are faults in the data.

An adaptive learning pigeon-inspired optimization based on mutation disturbance (ALPIO) is proposed for solving the problems of fuel consumption and threat avoidance in spacecraft cluster orbit reconstruction. First, considering the constraints of maintaining a safe distance between adjacent spacecraft within the spacecraft cluster and of avoiding space debris, the optimal performance index for orbital reconfiguration is proposed based on the fuel consumption required for path planning. Second, ALPIO is proposed to solve the path planning. Compared with traditional pigeon-inspired optimization, ALPIO uses the initialization of chaotic and elite backward learning to increase the population diversity, using a nonlinear weighting factor and adjustment factor to control the speed and accuracy of prepopulation convergence. The Cauchy mutation was implemented in the map and compass operator to prevent the population from falling into local optima, and the Gaussian mutation and variation factor were utilized in the landmark operator to prevent the population from stagnating in the late evolution. Through simulation experiments using nine test functions, ALPIO is shown to significantly improve accuracy when obtaining the optimum compared with PSO, PIO, and CGAPIO, and orbital reconfiguration consumes less total fuel. The trajectory of path planning for ALPIO is smoother than those of other optimization methods, and its obstacle avoidance path is the most stable.

Existing robust filters under generalized nonstationary noises conditions are difficult to choose suitable prior parameters, this brief proposed a Pearson-type VII distribution with adaptive parameters selection based interacting multiple model (IMM) Kalman filter (IMMPVII KF). In the model conditional filtering process, both the one-step prediction and the measurement likelihood are modeled as Pearson-type VII distributions. They are decomposed into Gaussian-Gamma Hierarchies (GGH), which are then matched to the time-varying heavy-tailed properties of the noises by pre-selecting the sets of shape and rate parameters and the variational Bayesian (VB) technique. Finally, a new model probability update method for filter under non-Gaussian conditions is derived. Simulation results show that the filter proposed in this brief has better robustness and adaptability to generalized non-stationary noises than the existing filters.

Spacecraft require a large-angle manoeuvre when performing agile manoeuvring tasks, therefore a control moment gyroscope (CMG) is employed to provide a strong moment. However, the control of the CMG system easily falls into singularity, which renders the actuator unable to output the required moment. To solve the singularity problem of CMGs, the control law design of a CMG system based on a cooperative game is proposed. First, the cooperative game model is constructed according to the quadratic programming problem, and the cooperative strategy is constructed. When the strategy falls into singularity, the weighting coefficient is introduced to carry out the strategy game to achieve the optimal strategy. In theory, it is proven that the cooperative game manipulation law of the CMG system converges, the sum of the CMG frame angular velocities is minimized, the energy consumption is small, and there is no output torque error. Then, the CMG group system is simulated. When the CMG system is near the singular point, it can quickly escape the singularity. When the CMG system falls into the singularity, it can also escape the singularity. Considering the optimization of angular momentum and energy consumption, the feasibility of the CMG system steering law based on a cooperative game is proven.

Estimating noise covariance matrices and suppressing outliers simultaneously is a huge challenge because they are coupled. In order to solve the nonlinear filtering problem with unknown non-Gaussian noise, the maximum correntropy criterion (MCC) is used to suppress the outliers, while the variational Bayesian (VB) technique is used to estimate the one-step prediction error covariance matrix (PECM) and the measurement noise covariance matrix (MNCM), and the nonlinear problem is solved by the third-order spherical radial cubature rule. In this paper, nominal measurement is constructed to eliminate the influence of outliers on the recursive estimation of MNCM. Through these operations, the variational Bayesian and maximum correntropy based cubature Kalman filter (VBMCCKF) and the maximum correntropy cubature Kalman filter (MCCKF) are derived. The interacting multiple model (IMM) fusion framework is used to promote the collaborative work of VBMCCKF and MCCKF, thus a novel filter is proposed in this paper. The simulation verifies the validity and universal applicability of the proposed filter. The estimation error of the proposed filter in spacecraft autonomous celestial navigation is about 30% less than that of the existing filter with the best performance.