Application of Reinforcement Learning to Generate Non-linear Optimal Feedback Controller for Ship's Automatic Berthing System

Implementation and Evaluation of Non-linear Optimal Feedback Control for Ship’s Automatic Berthing by Recurrent Neural Network

In this paper, the transient stabilization of power systems will be enhanced using optimal nonlinear feedback control. To this end, a partial differential equation called Hamilton-Jacobi-Belman (HJB), should be solved, which does not have a closed form solution in general. Nevertheless, there are some approximate solutions to solve HJB equations. In this paper, by using a Taylor Series expansion, a sub-optimal nonlinear control law is obtained. As a result, a transient controller for a single-machine infinite-bus power system is designed. Also, the superiority of the designed nonlinear controller in some grounds like increasing the region of asymptotic stability of the system or in other words, increasing the domain of validity is shown. Simulation results reveal the effectiveness of the proposed approach.

This paper presents the nonlinear closed-loop feedback control strategy for the spacecraft rendezvous problem with finite low thrust between libration orbits in the Sun–Earth system. The model of spacecraft rendezvous takes the perturbations in initial states, the actuator saturation limits, the measurement errors, and the external disturbance forces into consideration from an engineering point of view. The proposed nonlinear closed-loop feedback control strategy is not analytically explicit; rather, it is implemented by a rapid re-computation of the open-loop optimal control at each update instant. To guarantee the computational efficiency, a novel numerical algorithm for solving the open-loop optimal control is given. With the aid of the quasilinearization method, the open-loop optimal control problem is replaced successfully by a series of sparse symmetrical linear equations coupled with linear complementary problem, and the computational efficiency can be significantly increased. The numerical simulations of spacecraft rendezvous problems in the paper well demonstrate the robustness, high precision, and dominant real-time merits of the proposed closed-loop feedback control strategy.

The main transmission system of a scraper conveyor direct-driven by the high-power permanent magnet synchronous motor (PMSM) is taken as a study object. With the effect of the nonlinear friction torque caused by the nonuniformity of the transported coal quality in the operation process considered, the torsional vibration bifurcation mechanism and the corresponding control measures for the main transmission system of the scraper conveyor are investigated. Firstly, based on the Lagrange–Maxwell principle, the global electromechanical-coupling dynamic models for the main transmission system of the scraper conveyor are constructed. Secondly, by the Routh–Hurwitz stability criterion, the Hopf bifurcation characteristics of the main transmission system are analyzed to reveal the influence of supercritical bifurcation and subcritical bifurcation on the torsional oscillation of the transmission shafting. Thirdly, in order to suppress the system unstable oscillation caused by the Hopf bifurcation, the motor speed is fed back to construct the nonlinear state feedback controller for the quadrature axis current of the PMSM by the $$I_{d}=0$$ vector control strategy. Similarly, on the basis of the Routh–Hurwitz criterion, the influence of the linear feedback coefficient in the nonlinear state feedback controller on the system bifurcation position is discussed. Meanwhile, by the central manifold theory and canonical form theory, the effect of the square and cubic nonlinear feedback coefficients on the Hopf bifurcation type of the torsional vibration and the amplitude of the stable limit cycle are investigated. Finally, the numerical simulation results show the effectiveness of the designed controller.

Introduction Reinforcement learning formalizes the concept of learning from interactions.1 Broadly, reinforcement learning focuses on a setting in which an agent (decision maker) sequentially interacts with an environment that is partially unknown to them. At each stage, the agent takes an action and receives a reward. The objective of the agent is to maximize rewards accumulated in the long run. There are many situations in health care where decisions are made sequentially for which reinforcement learning approaches could prove useful for decision making. Throughout this article, we consider treatment prescription as an archetypical example to connect reinforcement learning concepts to a health care setting. In this setting, the care provider, the prescribed treatment, and the patients can be viewed as the agent, the action, and the environment, respectively, as depicted in Figure 1.Figure 1: Sequential treatment of AKI or CKD complications modeled as a reinforcement learning problem.Background In this section, with the objective of making reinforcement learning literature more accessible to a clinical audience, we briefly introduce related fundamental concepts and approaches. We refer the interested reader to Sutton and Barto1 for a comprehensive introduction to reinforcement learning. Markov Decision Processes Markov decision processes (MDPs) are a formalism of the sequential decision-making problem that has been central to the theoretical and practical advancements of reinforcement learning. In each stage of an MDP, the agent observes the state of the environment and takes an action, which, in turn, results in a change of the state. This change of state is assumed to be probabilistic with the next state being determined only by the preceding state, the chosen action, and the transition probability. The agent also receives a reward that is a function of the taken action, the preceding state, and the subsequent state. In an MDP, the objective of the agent is to maximize the return defined as the reward accumulated over a time horizon. In some applications, it is common to consider the horizon to be infinite, in which case the future rewards are discounted by a factor smaller than one. The selection of action by the agent on the basis of the observed state is known as the policy. More formally, a policy is a probabilistic mapping from states to each possible action. Because the policy and the reward are a function of the state, it is critical to estimate the utility of being in a certain state. More specifically, the value function is defined as the expected return starting from a given state under the chosen policy. Under this formalism, the objective of the agent is to find the optimal policy that maximizes the value function for all states. Reinforcement Learning Methods Action-value methods are a class of reinforcement learning methods in which the actions are chosen on the basis of the estimation of their long-term value. A prominent example of an action-value method is Q-learning in which the agent iteratively takes actions with the highest estimated values and updates the action-state value function on the basis of new observations. Policy gradient methods are another class of reinforcement learning methods that seek to optimize the policy directly instead of choosing actions on the basis of their respective estimated value. Such methods could be advantageous in health care applications that entail a large number of possible actions, e.g., when recommending a wide range of drug dosages or treatment options. Clinical Applications Reinforcement learning frameworks and methods are broadly applicable to clinical settings in which decisions are made sequentially. A prominent clinical application of reinforcement learning is for treatment recommendation, which has been studied across a variety of diseases and treatments including radiation and chemotherapy for cancer, brain stimulation for epilepsy, and treatment strategies for sepsis.2–5 In such treatment recommendation settings, a policy is commonly known as a dynamic treatment regime. There are various other clinical applications of reinforcement learning including diagnosis, medical imaging, and decision support tools (see refs. 2–5 and the references therein). Reinforcement Learning in Nephrology Although there have been recent applications of machine learning in nephrology,6,7 to the best of the authors' knowledge, the application of reinforcement learning to nephrology has been primarily limited to optimizing the erythropoietin dosage in hemodialysis patients.8,9 However, there are other settings where reinforcement learning has the potential to improve patient care in nephrology. For example, reinforcement learning methods can be adopted in the treatment of the complications of AKI or CKD (Figure 1). In this problem, the state models the conditions of the patient (e.g., vital signs, laboratory test results including urine and blood tests, and urine output measurements). The action refers to the treatment options (e.g., the dosage of medications such as sodium polystyrene sulfonate, and hemodialysis). The reward models the improvement in patient conditions. Similarly, reinforcement learning can help automate and optimize the dosage of immunosuppressive drugs in kidney transplants. Challenges and Opportunities Despite the success of reinforcement learning in several simplified clinical settings, their large-scale application to patient care faces several open challenges. The complexity of human biology complicates modeling clinical decision making as a reinforcement learning problem. The state space in such settings is often enormous, which could make a purely computational approach infeasible. Moreover, modeling all potential objectives a priori as a reward function may not be feasible. To overcome these challenges and realize the potential of reinforcement learning, clinical insight can play a pivotal role. More specifically, restricting the state space to only include highly relevant clinical variables could greatly reduce the computational complexity. Furthermore, using inverse reinforcement learning,2 relevant reward functions can be learned from retrospective studies assuming the optimality of clinical decisions. Another critical challenge is addressing moral and ethical concerns. It is imperative to ensure that reinforcement learning methods do not cause harm to the patient. To this end, there exists a need for a thorough validation of such methods before their use in patient care. Hence, there is a need to go beyond retrospective studies that have been used for the proof of concept of most existing reinforcement learning methods in health care applications.2,3 The lessons learned from the success of reinforcement learning in other application areas (e.g., self-driving cars) can help navigate the path to realizing its potential in health care. Accessible open-source simulation environments that enable researchers to compare various approaches are essential to the field of reinforcement learning. OpenAI Gym is currently the leading toolkit containing a wide range of simulated environments, e.g., surgical robotics.10 The development of high-quality and reliable simulation environments for nephrology and other health care applications can facilitate the development and validation of reinforcement learning methods beyond limited retrospective studies. The adoption of methods validated in such simulation environments in actual clinical settings will require clinicians' oversight. Similar to how self-driving cars require a human driver to ensure collision avoidance, clinicians' oversight is critical to ensure the safety of the patients, especially in the early stages of the adoption of reinforcement learning methods. The data from clinicians' decisions (e.g., overruling the automated treatment recommendation) can be used to improve the reliability of autonomous systems over time and reduce the burden of clinicians' oversight.

Deep reinforcement learning is the technology of artificial neural networks in the field of decision-making and control. The traditional model-free reinforcement learning algorithm requires a large amount of environment interactive data to iterate the algorithm. This model’s performance also suffers due to low utilization of training data, while the model-based reinforcement learning (MBRL) algorithm improves the efficiency of the data, MBRL locks into low prediction accuracy. Although MBRL can utilize the additional data generated by the dynamic model, a system dynamics model with low prediction accuracy will provide low-quality data and affect the algorithm’s final result. In this paper, based on the A3C (Asynchronous Advantage Actor-Critic) algorithm, an improved model-based deep reinforcement learning algorithm using a learning degree network (MBRL-LDN) is presented. By comparing the differences between the predicted states outputted by the proposed multidynamic model and the original predicted states, the learning degree of the system dynamics model is calculated. The learning degree represents the quality of the data generated by the dynamic model and is used to decide whether to continue to interact with the dynamic model during a particular episode. Thus, low-quality data will be discarded. The superiority of the proposed method is verified by conducting extensive contrast experiments.

High order optimal feedback control of space trajectories with bounded control

A EROSPACE engineering applications greatly stimulated the development of optimal control theory during the 1950s and 1960s, where the objective was to drive the system states in such a way that some defined cost was minimized. This turned out to have very useful applications in the design of regulators (where some steady state is to be maintained) and in tracking control strategies (where some predetermined state trajectory is to be followed). Among such applications was the problem of optimal flight trajectories for aircraft and space vehicles. Linear optimal control theory in particular has been very well documented and widely applied, where the plant that is controlled is assumed linear and the feedback controller is constrained to be linear with respect to its input. However, the availability of powerful low-cost microprocessors has spurred great advantages in the theory and applications of nonlinear control. The competitive era of rapid technological change, particularly in aerospace exploration, now demands stringent accuracy and cost requirements in nonlinear control systems. This has motivated the rapid development of nonlinear control theory for application to challenging, complex, dynamical real-world problems, particularly those that bear major practical significance in aerospace, marine, and defense industries. Infinite-time horizon nonlinear optimal control (ITHNOC) presents a viable option for synthesizing stabilizing controllers for nonlinear systems by making a state-input tradeoff, where the objective is to minimize the cost given by a performance index. The original theory of nonlinear optimal control dates from the 1960s. Various theoretical and practical aspects of the problem have been addressed in the literature over the decades since. In particular, the continuous-time nonlinear deterministic optimal control problem associated with autonomous (time-invariant) nonlinear regulator systems that are affine (linear) in the controls has been studied by many authors. The long-established theory of optimal control offers quite mature and well-documented techniques for solving this control-affine nonlinear optimization problem, based on dynamic programming or calculus of variations, but their application is generally a very tedious task. Bellman’s dynamic programming approach reduces to solving a nonlinear first-order partial differential equation (PDE), expressed by the Hamilton–Jacobi–Bellman (HJB) equation. The solution to the HJB equation gives the optimal performance/cost value (or storage) function and determines an optimal control in feedback form under some smoothness assumptions. Alternatively, in the classical calculus of variations, optimal control problems can be characterized locally in terms of the Hamiltonian dynamics arising from Pontryagin’s minimum principle. These are the characteristic equations of the HJB PDE, which result in a nonlinear, constrained two-point boundary value problem (TPBVP) that, in general, can only be solved by successive approximation of the optimal control input using iterative numerical techniques for each set of initial conditions. Numerically, even though the nonlinear TPBVP is somewhat easier to solve than the HJBPDE, control signals can only be determined offline and are thus best suited for feedforward control of plants for which the state trajectories are known a priori. Therefore, contrary to the dynamic programming approach, the resultant control law is not generally in feedback form. Open-loop control, however, is sensitive to random disturbances and requires that the initial state be on the optimal trajectory. In contrast, nonlinear optimal feedback has inherent robustness properties (inherent in the sense that it is obtained by ignoring uncertainty and disturbances). The potential difficulty with the HJB approach is that no efficient algorithm is available to solve the PDE when it is nonlinear and the problem dimension is high, making it impossible to derive exact expressions for optimal controls for most nontrivial problems of interest. The optimal can only be computed in special cases, such as linear dynamics and quadratic cost, or very low-dimensional systems. In particular, if the plant is linear time invariant (LTI) and the (infinite-time) performance index is quadratic, then the corresponding HJB equation for this infamous linear-quadratic regulator (LQR) problem reduces to an algebraic Riccati equation (ARE). Contrary to the well-developed and widely applied theory and computational tools for theRiccati equation (for example, see [1]), theHJB equation is difficult, if not impossible, to solve for most practical applications. The exact solution for the optimal control policies is very complex

Synchronization analysis of nabla fractional-order fuzzy neural networks with time delays via nonlinear feedback control

We propose an algorithm for nonlinear receding-horizon output feedback control by combining a real-time nonlinear optimal state feedback control algorithm and a real-time nonlinear optimal state estimation algorithm. The algorithms of state feedback control and state estimation are obtained by minimizing receding-horizon performance indexes. Application of the stabilized continuation method results in algorithms that do not involve any successive approximation methods. Simulation results of a 2-wheeled car clarify characteristics of the output feedback control algorithm.

Reinforcement Learning (RL) is a successful and established Artificial Intelligence (AI) method, particularly with recent groundbreaking progress in Deep Reinforcement Learning (DRL). Reinforcement learning is very well suited for sequential decision-making tasks, wherein a learned agent learns an optimal policy after many interactions with an environment. The present paper examines the application of reinforcement learning for automating screening of literature in academic research, particularly in the fields of computer science and e-learning. Keyword filtering techniques, while predominantly applied, are found to be inflexible as well as unable to capture the dynamic nature of research themes. To overcome such constraints, we present a Deep Q-Network (DQN)-based reinforcement learning model that combines reinforcement learning with the Semantic Scholar API to enhance research paper classification based on dynamically acquired decision rules. The proposed reinforcement learning model was trained and tested with a dataset of 8,934 research papers, accessed by systematic searching. The agent filtered and picked 11 effective papers depending on improved selection criteria like publication date, keyword relevance, and scholarly topic provided. The model iteratively optimizes the decision-making process through reward-based learning and therefore maximizes categorization accuracy over time. Test experiments demonstrate utilization of RL-based suggested framework yields classification accuracy at 91.5%, recall at 86.3%, and precision at 89.7%. A comparison test demonstrates that the approach performs 12.5% better on recall and 9.8% better on accuracy compared to traditional keyword-filtering methods. The finding confirms the power of the model in minimizing false positives and false negatives for screening literature, hence proving the scalability and adaptability of reinforcement learning in managing high academic data. This work offers a scalable, cognitive approach to conducting systematic reviews of literature through the application of reinforcement learning to programmatically execute work in academic research. The work shows the promise of reinforcement learning to further enhance research methodology, make literature reviews more effective, and facilitate more knowledgeable decision-making in fast-changing scientific disciplines. Further research will be focused on incorporating hybrid AI models with multi-agent systems of reinforcement learning for responsiveness and classification enhancements.

Finite-time partial stability and stabilization, and optimal feedback control

• Proposed a dynamic energy management system for hybrid sources in an electric vehicle. • Buck-boost H-bridge converters are introduced for bidirectional power flow control. • System non-linearities are addressed using a nonlinear state feedback controller. • The proposed system accurately delivers instantaneous vehicle speed and torque. This paper describes a methodology to control a multi-source battery-capacitor hybrid EV incorporating a dynamic power splitting strategy. Considering the requirement for precise speed control characteristics of the EV and the non-linear behavior of traction motor with system disturbances, a non-linear state feedback controller (NLSFC) with a disturbance observer is suggested as the speed control technique for the EV traction motor. To interface a parallel configured multi-source system, a non-inverted buck-boost H-bridge is proposed as the source side converter for effective recovery of the braking energy. Detailed control design for traction motor and converter is also presented for efficient control of EV. Besides, an adaptive energy management strategy considering power profile and dynamic source characteristics is incorporated into the overall control system. The simulation results from various case studies indicate that the proposed system can follow the required vehicle speed and torque despite the sudden change in EV load torque. Furthermore, the comparative analysis between NLSFC and PI controller shows improvement with steady-state error reduction from 0.43 to 0.0014 and peak overshoot drop from 12.07% to 3.51% during acceleration of the system when proposed NLSFC is adopted. The proposed solutions will be implemented in a prototype EV system in the future.

Chaotification of vibration isolation floating raft system via nonlinear time-delay feedback control

Reinforcement learning (RL), a promising framework for data-driven decision making in an uncertain environment, has successfully been applied in many real-world operation and control problems. However, the application of RL in a large-scale decentralized multi-agent environment remains a challenging problem due to the partial observability and limited communications between agents. In this paper, we develop a model-based kernel RL approach and a model-free deep RL approach for learning a decentralized, shared policy among homogeneous agents. By leveraging the strengths of both these methods, we further propose a novel deep ensemble multi-agent reinforcement learning (MARL) method that efficiently learns to arbitrate between the decisions of the local kernel-based RL model and the wider-reaching deep RL model. We validate the proposed deep ensemble method on a highly challenging real-world air traffic control problem, where the goal is to provide effective guidance to aircraft to avoid air traffic congestion, conflicting situations, and to improve arrival timeliness, by dynamically recommending adjustments of aircraft speeds in real-time. Extensive empirical results from an open-source air traffic management simulation model, developed by Eurocontrol and built on a real-world data set including thousands of aircrafts, demonstrate that our proposed deep ensemble MARL method significantly outperforms three state-of-the-art benchmark approaches.