This paper studies the finite-time trajectory tracking control of a space manipulator under model uncertainty, external disturbance and actuator saturation. The dynamics of space manipulator is derived using Kane's method. Considering the model uncertainty that may exist when the manipulator captures an unknown payload, a radial basis function (RBF) neural network (NN) is used to estimate the uncertain model of the space manipulator. An auxiliary system is designed to compensate the actuator saturation. Then a NN-based adaptive terminal sliding mode controller is proposed for trajectory tracking of a space manipulator. The stability of the proposed controller is analyzed using Lyapunov theory. Numerical simulations are conducted to verify the effectiveness of the proposed controller.

The Krasnoselskii-Pokrovskii (KP) model, as one of the popular operator-based hysteresis models, is commonly used to describe the hysteresis nonlinearities, especially in the smart materials-based actuators. Due to the complex formulation of the KP model, it is a great challenge to construct the inverse for the KP model. In this paper, an inverse multiplicative structure (IMS) is employed to find the inverse of the KP model. The merit of IMS is that this approach is simple to implement and the detailed knowledge of the hysteresis model is not required. However, the IMS technique cannot be directly applied to construct the inverse compensator for the KP model due to its complexity. Toward this problem, a new expression of the KP model is developed, where the input variable is expressed explicitly. With this new expression, the KP model can fit into IMS. Experiments are conducted to validate the effectiveness of the developed inverse compensator on a piezoelectric platform.

This paper studies the distributed adaptive control of a team of underactuated flexible spacecraft under leader-follower architecture with the measurements of the rigid bodies only. By treating the flexible spacecraft as an underactuated Lagrange system, an adaptive control strategy is proposed with the feedback of the generalized coordinates, velocities and accelerations of the rigid bodies. Then an extended state observer is introduced to design a distributed adaptive controller without the measurement of the generalized accelerations

This letter presents an approach for implementing game-theoretic decision-making in combination with deep reinforcement learning to allow vehicles to make decisions at an unsignalized intersection by use of 2D Lidar to obtain their observations of the environment. The main novelty of this work is modeling multiple vehicles in a complex interaction scenario simultaneously as decision-makers with conservative, aggressive, and adaptive driving behaviors. The game model allows anticipating reactions of additional vehicles to the movements of the ego-vehicle without using any specific coordination or vehicle-to-vehicle communication. The solution of the game is based on cognitive hierarchy reasoning and it uses a deep reinforcement learning algorithm to obtain a near-optimal policy towards a specific goal in a realistic simulator (ROS-Gazebo). The trained models have been successfully tested on the simulator after training. Experiments show that the performance of the lab cars in the real-world is consistent with it in the simulation environment, which may have great significance to improve the safety of self-driving cars, as well as may reduce their dependence on road tests.

This article studies fully distributed formation control problems for multiquadrotor systems with external disturbances via an event-triggered approach. For leader–follower multiquadrotor systems, a fully distributed event-triggered formation control framework is proposed by utilizing an adaptive sliding-mode control approach, which does not rely on any global information of the network topology. The finite-time reachability of the sliding-mode surface can be guaranteed for the states of the nonlinear, coupled, and underactuated system subject to external disturbances. Meanwhile, the zeno behavior of the proposed event-triggered protocol is proved to be excluded. A novel dynamic sliding-mode surface is designed to guarantee the formation performance as the quadrotor state trajectories move on the constructed sliding manifold. Via Lyapunov stability theory, sufficient conditions to ensure the formation results are derived for multiquadrotor systems. Simulations and experiments are conducted to validate the effectiveness of the proposed control scheme.

Multiple flexible manipulators can be used to complete some repeatable missions. Each flexible manipulator can be described as an under-actuated Lagrangian system based on the assumed modes method. Also, the actuator nonlinearity may deteriorate the system performance. Hence, this paper aims to develop a distributed controller to solve the leader-follower consensus of multiple flexible manipulators with uncertain parameters, unknown disturbances and actuator dead zones. The disturbances are classified as repeatable and non-repeatable ones. The adaptive, iterative learning and sliding mode control techniques are used to handle uncertain parameters, repeatable and non-repeatable disturbances, respectively. Based on a dead-zone inverse and a finite-time observer, a distributed controller is developed to drive the flexible manipulators to track a moving leader and keep the flexible vibrations bounded simultaneously. Experimental results are presented to verify the effectiveness of the proposed controller.