Abstract Wind power plays a vital role in the global effort towards net zero. A recent figure shows that 93GW new wind capacity was installed worldwide 2020, leading to 53% year-on-year increase. The control system is core of farm operations and has an essential influence on farm’s capture efficiency, economic profitability, operation maintenance cost. However, inherent complexities farms aerodynamic interactions among turbines cause significant barriers design. industry recognized...

This paper addresses the torque and pitch control problems of wind turbines. The main contribution this work is development an innovative reinforcement learning (RL)-based method targeting turbine applications. Our RL-based framework synergistically combines advantages deep neural networks (DNNs) model predictive (MPC) technologies. proposed strategy data-driven, adapting to real-time changes in system dynamics enhancing performance robustness. Additionally, incorporation MPC structure...

This work proposes an innovative path-following control method, anchored in deep reinforcement learning (DRL), for unmanned underwater vehicles (UUVs). approach is driven by several new designs, all of which aim to enhance efficiency and effectiveness achieve high-performance UUV control. Specifically, a novel experience replay strategy designed integrated within the twin-delayed deterministic policy gradient algorithm (TD3). It distinguishes significance stored transitions making trade-off...

This paper addresses the translational control problem for final phase of spacecraft rendezvous and docking. For safety concerns, during approach process, pursuer is required to strictly comply with approaching path constraints should also have obstacle-avoidance ability at same time. A novel potential function employed describe these requirements. Two distinct constrained zones are designed applications them analyzed as well. Then, based on a time-varying sliding manifold, an adaptive law...

This paper addresses the integrated attitude and position control problem for final phase proximity operations of spacecraft autonomous rendezvous docking, in which important motion constraints chaser are considered. On one hand, to ensure reliable real-time measurements relative information between two spacecraft, relevant sensor system is required continuously point toward target; on other safety concerns, also needs follow a specified approach path constraint. A special...

This article addresses the tracking control problem for a class of nonlinear systems described by Euler–Lagrange equations with uncertain system parameters. The proposed scheme is capable guaranteeing prescribed performance from two aspects: 1) special parameter estimator prescribed-performance properties embedded in scheme. not only ensures exponential convergence estimation errors under relaxed excitation conditions but also can restrict all estimates to predetermined bounds during whole...

Results are presented for a study of dual-quaternion-based fault-tolerant control spacecraft tracking. First, six-degrees-of-freedom dynamic model under description is employed to describe the relative coupled motion target-pursuer tracking system. Then, novel method proposed enable pursuer track attitude and position target even though its actuators have multiple faults. Furthermore, based on time-varying sliding manifold, finite-time stability closed-loop system theoretically guaranteed,...

This article addresses the attitude reorientation problems of rigid bodies under multiple state constraints. A novel reinforcement learning (RL)-based approximate optimal control method is proposed to make tradeoff between cost and performance. The novelty lies in that it guarantees constraint handling abilities on forbidden zones angular velocity limits. To achieve this, barrier functions are employed encode information into function. Then, an RL-based strategy developed function policy....

This article is devoted to real-time optimal attitude reorientation control of rigid spacecraft control. Particularly, two typical practical problems—actuator misalignment and forbidden pointing constraints are considered. Within the framework adaptive dynamic programming (ADP), a novel constrained scheme proposed.In this design, special reward function developed characterize environment feedback deal with constraints. Notably, argument term introduced for overcoming inevitable difficulty in...

This article addresses the crucial requirements in spacecraft attitude control: prescribed performance guarantees under actuator saturation and real-time cost optimization. As an application-oriented study, approximate optimal control scheme is proposed for this objective. To be specific, constraint converted into system dynamics merged adaptive dynamic programming design philosophy. Subsequently, online learning law designed based on a special saturated Hamilton–Jacobi–Bellman error, which...

The high system complexity and strong wake effects bring significant challenges to wind farm operations. Conventional control methods may lead degraded power generation efficiency. A reinforcement learning (RL)-based approach is proposed in this paper handle these issues, which can increase the long-term farm-level subject while without requiring analytical models. method significantly distinct from existing RL-based approaches, whose computational complexities usually heavily with of total...

A model-free deep reinforcement learning (DRL) method is proposed in this article to maximize the total power generation of wind farms through combination induction control and yaw control. Specifically, a novel double-network (DN)-based DRL approach designed generate policies for thrust coefficients angles simultaneously separately. Two sets critic-actor networks are constructed end. They linked by central power-related reward, providing coordinated structure while inheriting mechanism's...

In this article, a deep reinforcement learning (RL)-based control approach with enhanced efficiency and effectiveness is proposed to address the wind farm problem. Specifically, novel composite experience replay (CER) strategy designed embedded in deterministic policy gradient (DDPG) algorithm. CER provides new sampling scheme that can mine information of stored transitions in-depth by making tradeoff between rewards temporal difference (TD) errors. Modified importance-sampling weights are...

This article aims to address the wind-farm power tracking problem, which requires farm's total generation track time-varying references and, therefore, allows wind farm participate in ancillary services such as frequency regulation. A novel preview-based robust deep reinforcement learning (PR-DRL) method is proposed handle tasks are subject uncertain environmental conditions and strong aerodynamic interactions among turbines. To our knowledge, this for first time that a data-driven...

This paper addresses the attitude-maneuver control problem for a rigid-body spacecraft in presence of attitude-constrained zones as well input saturation. More specifically, are properly encoded, and an admissible artificial potential function (APF) is developed under unit-quaternion representation. The elaborately designed APF ensures unique minimum without requiring convexity constrained zones, which yields less stringent condition from previous approaches. Benefiting bounded property...

This article proposes areinforcement learning (RL)-based six-degree-of-freedom (6-DOF) control scheme for the final-phase proximity operations of spacecraft. The main novelty proposed method are from two aspects: 1) closed-loop performance can be improved in real-time through RL technique, achieving an online approximate optimal subject to full 6-DOF nonlinear dynamics spacecraft; 2) nontrivial motion constraints considered and strictly obeyed during whole process. As a stepping stone,...

A new disturbance observer-based control method is presented in this paper to address the attitude tracking problem of rigid-body spacecraft presence external disturbances and parameter uncertainties. Particularly, a sliding mode observer (SMDO) designed. The most important feature SMDO relaxation assumption that must be constants or changing at slow rate, which typical required these classes problems concerning (DO) design but hard guarantee from standpoint practical engineering. In...

This brief proposes a novel data-driven control scheme to maximize the total power output of wind farms subject strong aerodynamic interactions among turbines. The proposed method is model-free and has robustness, adaptability, applicability. Particularly, distinct from state-of-the-art farm methods that commonly use steady-state or time-averaged data (such as turbines' outputs under steady conditions models) carry out learning, directly mines in-depth time-series measured at turbine rotors...

In this paper, a novel adaptive dynamic programming (ADP)-based optimal control method is developed for discrete-time systems subject to constraints and disturbances. Particularly, safe policy iteration scheme designed handle state input constraints, including both hard soft by converting the original improvement strategy into constrained optimization problem with prescribed cost function. After that, an actor-critic-disturbance framework introduced address problem. The robust safety against...

Based on the dual-quaternion description, a smooth six-degree-of-freedom observer is proposed to estimate incorporating linear (translational) and angular velocity, called dual-angular for rigid bodies. To establish observer, some important properties of dual vectors quaternions are presented proved, additionally, kinematics dualtransformation matrices deduced, transition relationship between transformation subsequently analyzed. An feature that all estimation states ensured be C <sup...