Continuous aerial sliding contact manipulation is a promising application of uncrewed aerial manipulators (UAMs). Existing studies have demonstrated such tasks on structured geometries, such as planes and cylindrical surfaces. However, unstructured contact targets introduce continuously varying surface geometry and coupled task constraints. These constraints form implicit manifolds in the planning space, making efficient planning a fundamental challenge for unstructured surface manipulation. To address this problem, this study proposes a perception–planning–control framework centered on a metric-consistent parameterized tree (MCPT*) planner. MCPT* transforms the task-constrained planning problem from the high-dimensional configuration space into a lower dimensional parameter space. It then uses a metric-consistent mapping to search for high-quality solutions. This spatial compression and metric transformation improve planning efficiency while preserving solution quality. By further combining an admittance filter with a robust controller based on a wrench estimator, the proposed framework maintains stable and continuous sliding contact during execution. Simulations and real-world experiments with a fully actuated UAM validate the planning efficiency, trajectory tracking performance, and continuous contact maintenance capability of the proposed framework.
The four-wheel independent steering (4WIS) system features redundant steering actuators and multiple operable modes, yet ensuring vehicle safety through mode switching upon an actuator fault remains a critical challenge. This article proposes a multiple-mode self-switching fault-tolerant control strategy to address single angle module fault in 4WIS vehicles. The strategy comprises a mode switching layer and a trajectory tracking layer. The mode switching layer employs a Lyapunov robust Hinf controller (LRHinfC). Its design is based on a mode-dependent Lyapunov–Krasovskii functional incorporating a prescribed exponential decay rate, which rigorously guarantees stability and robust performance against system delays and uncertainties. This controller ensures the smooth, coordinated zeroing of the faulty steering module and accurate angle tracking of the healthy modules, thereby achieving stable steering mode transitions. The trajectory tracking layer implements an adaptive multiple-order sliding mode controller. It introduces a composite sliding surface to suppress disturbances and employs an adaptive law with a clear proportional-leakage mechanism to estimate and compensate for the lumped uncertainties arising from faults and model mismatches. This enables consistent trajectory tracking performance across different, postfault steering modes. Based on the collaborative effect of the mode switching layer and trajectory tracking layer, ensuring the tracking performance of intelligent vehicles.
Wing attitude of hybrid-driven underwater gliders (HUGs) is a critical factor for hydrodynamic adjustment and navigation mode switching. However, existing controllable wing mechanisms typically rely on modifying wing geometric parameters for single-mode optimization, inherently limiting multimode adaptability. Inspired by the fin motion of the sunfish, this article proposes an HUG with controllable dihedral angle wings (HUG-CDAW), which adopts internal ring gear transmission to realize a functional reconfiguration. Specifically, this design enables seamless switching between distinct functional roles to significantly enhance depth-fixed navigation accuracy, steering capability, and roll angle stability, while preserving the streamlined axisymmetric profile. A dynamic model accounting for dihedral-dependent hydrodynamic coefficients is established and validated. Experiments demonstrate that the HUG-CDAW achieves a turning radius of 1.64 m (1.17 body lengths), representing a 54.9% reduction compared with the conventional movable battery steering. Due to the low inertia of the neutrally buoyant wings, the actuation energy is reduced by 62% and the response time by 95%. This article provides a new design strategy for advancing the multimode adaptability of underwater vehicles.
Integrating multiple shunted piezoelectric elements onto a structure has been widely recognized as an effective method for achieving broadband vibration control due to its lightweight and tunable features. However, the intricate electromechanical coupling induced by these patches presents a significant challenge to the optimal design of the associated shunt circuits. Furthermore, reliably predicting the vibration control performance prior to physical implementation remains elusive. To address these limitations, this study proposes an inverse design methodology through analytically deriving the optimal shunt circuit parameters for an arbitrarily given structure under light-damping conditions, thereby achieving user-defined vibration suppression targets. In particular, a coupling analysis framework underpinned by the associated simplified models is developed within the space spanned by the excitation-dependent representative basis, which can be extracted from experimentally measured response of the primary structure. This approach effectively avoids the computational complexity of the conventional model-based dynamic analysis while retaining the essential electromechanical interaction characteristics of the fully coupled system. The proposed inverse design strategy allows for achieving performance-driven vibration control, as demonstrated through numerical simulations and further validated experimentally. The method offers an efficient, predictive, and customized design pathway for multiresonant vibration control.
Steer-by-wire (SBW) systems are critical for vehicle safety and autonomous driving, heavily relying on precise angle tracking performance. While existing control frameworks primarily enhance robustness against mechanical disturbances, they largely overlook the actuator motor torque deviations induced by electromagnetic disturbances. To bridge this gap, this article proposes a novel robust angle tracking control framework featuring active actuator torque compensation (ATC). Specifically, an $H_{\infty }$ controller with decoupled angular and angular velocity loops is synthesized to effectively reject mechanical disturbances. Building upon this, an extended state observer (ESO) is designed to estimate and compensate for electromagnetic torque deviations in real time. This dual-layer architecture ensures both the generation of optimal torque commands and their highly accurate physical execution. Comprehensive simulation and vehicle test results demonstrate that the proposed controller significantly enhances tracking accuracy and reduces response latency.
