This paper proposes a method for designing disturbance observer (DOB) in vibration suppression control of two-inertia system. The key to a vibration suppression control of two-inertia system is eliminating the effect of vibration modes of the plant. For this purpose, feedback controller is designed to cancel the stable vibration modes of plant. However, this cancellation fails due to model error from parameter variations. In this paper, the DOB is employed to compensate for the effect of parameter variations. The Q-filter of DOB is designed with the following specifications: optimal disturbance suppression performance with restricted cutoff frequency; robust stability of closed-loop system against modeling error between the plant and its nominal model; structural restriction of Q-filter itself such as relative order and internal model order. This paper proposes a systematic and straightforward method to attain the above design specifications using standard H _∞ control framework. Experimental results verify that the proposed DOB can improve the transient response and effectively suppress the vibration and improve external disturbance removing performance as well.

This brief proposes a disturbance rejection control strategy for attitude tracking of an aircraft with both internal uncertainties and external disturbances. The proposed control strategy consists of a robust disturbance observer (DOB) and a nonlinear feedback controller. Specifically, a robust DOB is proposed to compensate the uncertain rotational dynamics into a nominal plant, based on which a nonlinear feedback controller is implemented for desired tracking performance. We first divide the practical rotational dynamics into the nominal part, external disturbances, and equivalent internal disturbances. Then, property of equivalent internal disturbances is explored for stability analysis. A robust DOB is optimized based on H _∞ theory to guarantee disturbance rejection performance and robustness against system uncertainties. A practical nonlinear feedback controller is hence applied to stabilize the compensated system based on backstepping approach. Experiments on a quadrotor testbed show that the proposed robust DOB can suppress the external disturbances and measurement noise, with the robustness against system uncertainties.

This brief proposes a method for designing a disturbance observer (DOB) to decouple joint interactions in robot dynamics with nonlinearity. The traditional DOB based on filter design theory has limited performance since the cut-off frequency of its Q-filter is the only tunable parameter to deal with disturbance suppression and model uncertainty. In this brief, a robust optimal design method is developed for the DOB, which can achieve optimal performance of suppressing disturbance by systematically shaping its Q-filter. Simulation results of application to a two-link manipulator with flexible joints show the improvements in disturbance suppression, which illustrates the validity of the proposed method.

Recent years have witnessed the popularity of integrating Siamese network into RGBT tracking for fast-tracking. However, these trackers mostly utilize the feature information of the last output layer and ignore the benefits of multi-layer information. In addition, they often adopt feature-level fusion for different modalities but fail to explore the strength of decision-level fusion, which may easily decrease their flexibility and independence. In this paper, a novel multi-layer attention aggregation Siamese network on the decision level is proposed for robust RGBT tracking. To be specific, a hierarchical channel attention Siamese network is built to recalibrate the extracted multi-layer features from RGB and thermal infrared images. This can focus on more discriminative features to learn robust feature representation. Then, a depth-wise correlation operation is performed to produce RGB and thermal response maps, respectively. To better exploit and utilize the complementary RGB and thermal information, a contribution-aware aggregation network is designed to adaptively aggregate them. Lastly, a classification and regression network is adopted to complete the bounding box prediction. Extensive experiments on four large-scale RGBT benchmarks demonstrate outstanding tracking ability over other state-of-the-art trackers.