Gaussian filters with deterministic sample points, such as the Unscented Kalman Filter (UKF), Cubature Kalman Filter (CKF), Gauss– Hermite Quadrature Filter (GHQF), etc., have been widely employed for nonlinear state estimation. However, these filters utilize a fixed set of sample points irrespective of the system's nonlinearity. While various studies have explored data-driven approaches to optimize the three parameters of the UKF, these methods do not generalize well to other Gaussian filters with a greater number of sample points. In this letter, we propose a novel neural network-based unified framework for finetuning sample points across all Gaussian filters with deterministic sample points. Specifically, we first pretrain a Multi-Layer Perceptron (MLP) to approximate the mapping from the state's mean and covariance to the sample points of the original Gaussian filter. The MLP then replaces the sample points generation strategy in the Gaussian filter and is further refined by maximizing the marginal likelihood of the observed measurement data. Simulation results demonstrate that the Gaussian filters leveraging the well-trained MLP as the sample point generation strategy achieve higher filtering accuracy compared to their original counterparts.

High-precision and safety control in face of disturbances and uncertainties is a challenging issue of both theoretical and practical importance. In this article, new adaptive anti-disturbance control schemes are proposed for a class of uncertain nonlinear systems with composite disturbances, including additive disturbances, multiplicative actuator faults, and implicit disturbances deeply coupled with system states. Both the cases with known and unknown control/fault directions are investigated. By properly fusing the techniques of disturbance observers and adaptive compensation, it is shown that all closed-loop signals are globally uniformly bounded and the tracking error converges to zero asymptotically, no matter the control/fault directions are known or not. In the case of known directions, the proposed control scheme, for the first time, guarantees asymptotic tracking andL$_{\infty}$tracking performance simultaneously in face of disturbances and actuator faults. Moreover, novel Nussbaum functions and a contradiction argument are introduced, which allow the system to have multiple unknown nonidentical control directions and unknown time-varying fault direction. Simulation results illustrate the effectiveness of the proposed control schemes.

The safety and high-precision attitude control of launch vehicles are threatened by degradation fault and multiple disturbances (such as model uncertainty, uncertain inertia, and external disturbance) during the reentry stage. To address these challenges, an adaptive sliding mode observer (ASMO)-based composite control scheme is proposed in this article for launch vehicles to achieve simultaneous compensation and suppression of the degradation fault and multiple disturbances. Since the mismatched model uncertainty that coupled with the system state exhibits strong uncertainty, an ASMO is designed to estimate it by adaptively learning the upper bound of the derivative of the mismatched model uncertainty. In order to attenuate the effect of the degradation fault, uncertain inertia, and external disturbance, three adaptive laws are accordingly designed to identify them online. By combining the ASMO and the designed adaptive laws, a composite controller is constructed, and the degradation fault and multiple disturbances are simultaneously compensated and suppressed. The coordinated optimization performance and refinement of antidisturbance control and fault-tolerant control are effectively enhanced. Moreover, by introducing a prescribed performance function, the attitude tracking error response is constrained within a predefined range. Simulation and experiments validate the effectiveness of the proposed scheme.

Depth estimation problem for the perspective vision systems has been extensively studied in the literature and the depth estimation can be achieved under the assumption that the exact camera motion parameters are available. However, in practice, it is hard to obtain the camera motion parameters exactly. For instance, the exact camera velocity is always unavailable and instead only the roughly estimated one can be obtained. This introduces the significant difficulties to achieve the depth estimation with partially uncertain camera motion parameters. In this article, we consider the depth estimation problem for the perspective vision system in the case that the camera angular velocities are accurate, while partial camera linear velocities are contaminated by some disturbances. This problem is theoretically formulated and solved for the first time in this article by proposing a new depth observer in the presence of the partially uncertain camera linear velocities. Local exponential convergence of the depth and disturbance estimates is achieved in presence of the partially uncertain camera linear velocities such that the estimation errors of states and disturbances converge to zero for constant disturbances or to a small error bound for bound time-varying disturbances. Simulations and experiments are carried out to verify the performance of the proposed observer.

The proportional-integral (PI) controller has proven to be a remarkably effective control strategy in the servo field. However, it is difficult for PI to handle the problem of multiple disturbances degrading the velocity-tracking capability of the gimbal system in the control moment gyro (CMG). In this article, we propose a solution to this problem while preserving the advantages of PI from an engineering standpoint. Starting from the refined disturbance analysis based on a 5 Nms CMG gimbal system experimental tests, multiple disturbances of the gimbal system can be classified according to frequency distribution in terms of 0$\sim $5 and 128 Hz. Subsequently, under the composite hierarchical anti-disturbance control frame, a peak filter and a disturbance observer addition improve the anti-disturbance capability of the PI-based gimbal system at the frequencies of 0$\sim $5 and 128 Hz. Notably, the proposed method has a relatively low order, which is appropriate for the processing power limit. Finally, the experimental results reveal that compared to three effective control schemes (PI, extended state observer-based composite controller, and PI-resonant controller), the proposed method can strikingly enhance the anti-disturbance ability and velocity-tracking performance of the gimbal system.

In this paper, a composite velocity-tracking control scheme is presented for improving dynamics response and control accuracy of the flexible gimbal system in the control moment gyro (CMG). Specifically, an${H}_{\infty } $optimal vibration controller (HOVC) is developed to weigh multi-performance of the gimbal system, such as stability, fast-tracking performance, and attenuation ability of multi-frequency-band disturbances. Meanwhile, a robust disturbance observer (RDO) designed on optimization technology is utilized to improve the nonlinear friction rejection ability of the gimbal system. Moreover, rigorous theoretical analysis based on the small-gain theory is provided to demonstrate the stability of the closed-loop gimbal system under multiple disturbances. Finally, experimental studies are presented to verify the effectiveness of the proposed method.