This paper focuses on the trajectory tracking control of unmanned underwater vehicles (UUVs) in the presence of dynamic uncertainties and time-varying external disturbances. Two adaptive integral terminal sliding mode control schemes, namely, adaptive integral terminal sliding mode control (AITSMC) scheme and adaptive fast integral terminal sliding mode control (AFITSMC) scheme are proposed for UUVs based on integral terminal sliding mode (ITSM) and fast ITSM (FITSM), respectively. Each control scheme is double-looped: composed of a kinematic controller and a dynamic controller. First, a kinematic controller is designed for each of the two control schemes. The two kinematic controllers are based on ITSM and FITSM, respectively. These kinematic controllers yield local finite-time convergence of the position tracking errors to zero meanwhile avoid the singularity problem in the conventional terminal sliding mode control (TSMC). Then, using the output of the kinematic controller as a reference velocity command, a dynamic controller is developed for each of the two control schemes. The two dynamic controllers are also based on ITSM and FITSM, respectively. An adaptive mechanism is introduced to estimate the unknown parameters of the upper bound of the lumped system uncertainty which consists of dynamic uncertainties and time-varying external disturbances so that the prior knowledge of the upper bound of the lumped system uncertainty is not required. The estimated parameters are then used as controller parameters to eliminate the effects of the lumped system uncertainty. The convergence rate of the integral terminal sliding variable vector is investigated and the local finite-time convergence of the velocity tracking errors to zero in the ITSM or FITSM is obtained. At last, based on the designed kinematic and dynamic controllers, the finite-time stability of the full closed-loop cascaded system is shown. The two proposed control schemes improve the tracking accuracy over the existing globally finite-time stable tracking control (GFTSTC) and adaptive nonsingular TSMC schemes, and enhance the robustness against parameter uncertainties and external disturbances over the GFTSTC scheme. Compared with the conventional adaptive integral sliding mode control (AISMC) scheme, the two proposed control schemes offer faster convergence rate and stronger robustness against dynamic uncertainties and time-varying external disturbances for the trajectory tracking control of UUVs due to involving the fractional integrator. Comparative numerical simulations are performed on the dynamic model of the Omni Directional Intelligent Navigator UUV for two trajectory tracking cases. The convergence rate and robustness to uncertainties and disturbances are quantified as the convergent time and bounds of the steady-state position and velocity tracking errors, respectively. The results show that the two proposed control schemes improve at least 20s in convergence rate and enhance about 2% robustness in position tracking and 20% robustness in velocity tracking over the AISMC scheme.

This paper focuses on the design of an adaptive second-order fast nonsingular terminal sliding mode control (ASOFNTSMC) scheme for the trajectory tracking of fully actuated autonomous underwater vehicles (AUVs) in the presence of dynamic uncertainties and time-varying external disturbances. First, a second-order fast nonsingular terminal sliding mode (SOFNTSM) manifold is designed to achieve a faster convergence rate than the existing second-order nonsingular terminal sliding mode (SONTSM) manifold. Then, by using this SOFNTSM manifold, an ASOFNTSMC scheme is developed for the fully actuated AUVs to track the desired trajectory. The designed SOFNTSM manifold yields local exponential convergence of the position and attitude tracking errors to zero, and the developed ASOFNTSMC scheme ensures that the error trajectories always move toward the SOFNTSM manifold and once they hit the manifold, remain on it in the presence of dynamic uncertainties and time-varying external disturbances. By deriving the expression of the bounding function of the system uncertainty and using adaptive technique to estimate the unknown parameters of the system uncertainty bounds, the ASOFNTSMC scheme does not require the prior knowledge of the upper bound of the system uncertainty. Meanwhile, through involving the discontinuous sign function into the time derivative of the control input, the ASOFNTSMC scheme eliminates the chattering without reducing the tracking precision. Compared with the existing adaptive SONTSM control (ASONTSMC) scheme, the proposed ASOFNTSMC scheme offers a faster convergence rate for the trajectory tracking control of fully actuated AUVs. Two comparative simulation cases performed respectively on two fully actuated AUVs demonstrate the superiority of the ASOFNTSMC scheme over the ASONTSMC scheme.

Collision-free space detection is a critical component of autonomous vehicle perception. The state-of-the-art algorithms are typically based on supervised deep learning. Their performance is dependent on the quality and amount of labeled training data. It remains an open challenge to train deep convolutional neural networks (DCNNs) using only a small quantity of training samples. Therefore, in this article, we mainly explore an effective training data augmentation approach that can be employed to improve the overall DCNN performance, when additional images captured from different views are available. Due to the fact that the pixels in collision-free space (generally regarded as a planar surface) between two images, captured from different views, can be associated using a homography matrix, the target image can be transformed into the reference view. This provides a simple but effective way to generate training data from additional multiview images. Extensive experimental results, conducted with six state-of-the-art semantic segmentation DCNNs on three datasets, validate the effectiveness of the proposed method for enhancing collision-free space detection performance. When validated on the KITTI road benchmark, our approach provides the best results, compared with other state-of-the-art stereo vision-based collision-free space detection approaches.

In leader-following formation, the leader plays a central role and is more vulnerable to malicious attacks. To effectively protect the leader, this article investigates an event-based formation control problem of autonomous surface vehicles (ASVs) with multiple attackers and actuator failure. A novel adaptive formation keeping and interception scheme is developed for ASVs. In this scheme, a formation keeping controller and an interception controller are designed under a unified framework by sharing same event-triggered mechanism and fault-tolerant structure. To bridge these two controllers, a degree and distance driven formation interception ( $\text{D}^3$ FI) strategy is designed by generating defenders. It is shown that with the proposed scheme, all attackers are intercepted without affecting formation keeping of remaining ASVs if the original topology is connected. Advantages of our scheme are: 1) stable controller switching is ensured during the formation cooperation and 2) the fault-tolerance is obtained with lower actuator updating frequencies and least number of adaptive parameters for each ASV.