This article investigates the problem of robust trajectory design for low-thrust spacecraft subject to state and observation uncertainties. An adaptive latent reinforcement learning (RL) scheme based on sequential latent variable models (SLVMs) is proposed to address this issue. First, an SLVM is employed for the representation learning of uncertain environments and for predicting future observations. Subsequently, by integrating representation learning based on the SLVM with proximal policy optimization (PPO), a stochastic latent PPO (SLPPO) scheme is introduced. Distinct from existing methods, the control policy is derived from learned stochastic latent variables rather than raw uncertain observations, which effectively mitigates the adverse impact of uncertainties on control performance. Furthermore, to enhance training efficiency, an improved dense reward shaping mechanism is designed based on the observation predictions from the SLVM and adaptive techniques. Finally, numerical simulations of two rendezvous missions validate the effectiveness of the proposed approach.

This paper investigates a distributed model predictive control (DMPC) for linear heterogeneous systems tracking arbitrary periodic references. The control objective consists of two parts: (i) driving the output of each subsystem consensus; (ii) steering the outputs as close as possible to an exogenous periodic reference. The artificial state reference and control input are considered as decision variables to track unreachable references. The optimal control problem (OCP) is then solved in a distributed manner using Alternating Direction Multiplier Method (ADMM). The proposed method does not need ADMM convergence at each time step, which greatly reduces the computation time. Under several mild assumptions, the feasibility of the OCP and the closed-loop asymptotic stability with respect to an optimal reachable cooperative trajectory are presented. The performance of the approach is demonstrated with some simulation results.Note to Practitioners—The paper is motivated by the problem of cooperative tracking unreachable references for heterogeneous systems. The generation of the reference signal often ignores the dynamics feature of systems, leading to such reference may not be fully tracked (unreachable reference). However, existing methods either lack optimality or cannot achieve cooperative tracking of unreachable references. Therefore, this study develops a novel DMPC approach to make up for the above lack. In addition, the proposed method greatly reduces the computation time while ensuring optimality, and has a wider initial feasibility. The proposed controller can be extended to cooperative track unreachable constant signal. The proposed method can be used for highly collaborative tasks, such as formation missions, collaborative transportation and spacecraft collaboration. In future research, we will address the problem of cooperative tracking unreachable references for nonlinear heterogeneous systems.

This paper addresses the challenges of energy consumption, driving safety, and robustness in tracking control for autonomous vehicles under stochastic disturbances by proposing a robust economic model predictive control (REMPC) algorithm without terminal constraints. The disturbances considered include high-probability small disturbances as well as low-probability large disturbances in tracking control. A tightening constraint is introduced to ensure robustness against small disturbances, leveraging constraint tightening theory. Additionally, a maximum probability interval is derived to account for the large disturbances that the vehicle can withstand. To further enhance driving safety and energy efficiency, a variable space constraint is adaptively designed based on road slope information. The paper demonstrates recursive feasibility and robust asymptotic performance of the optimization problem with variable space constraints. Robust asymptotic stability in probability of the system is ensured by deriving a probability condition and a lower bound on the prediction horizon, with an exhaustive principle for selecting the prediction horizon. The efficiency and superiority of the proposed REMPC algorithm are verified by a comprehensive case study under various operating conditions.

This paper presents a hierarchical economic model predictive control (EMPC) framework for electric vehicles (EVs) to address traffic congestion, energy consumption, and driving safety in tracking control. The framework comprises a higher-level planner and a lower-level controller that act in tandem to achieve driving safety and economic efficiency while ensuring control performance. The higher-level planner employs an event-triggered logic by taking traffic density and economic consumption into account. The resulting optimization problem is only solved to update the reference control input of lower-level controller when the traffic density violates a pre-specified threshold. The lower-level controller ensures driving safety and control performance of EVs by designing adaptive inter-vehicle distance constraints. In addition, the robustness of the EMPC framework is ensured by adopting a robust constraint tightening policy. Recursive feasibility analyses of the optimization problems in both levels of the framework are also conducted. Rigorous proofs of asymptotic average performance and stability analysis are guaranteed for the closed-loop system. The proposed hierarchical EMPC algorithm is demonstrated to be effective and superior in a case study.

Designing an efficient learning-based model predictive control (MPC) framework for ducted-fan unmanned aerial vehicles (DFUAVs) is a difficult task due to several factors involving uncertain dynamics, coupled motion, and unorthodox aerodynamic configuration. Existing control techniques are either developed from largely known physics-informed models or are made for specific goals. In this regard, this article proposes a compound learning-based MPC approach for DFUAVs to construct a suitable framework that exhibits efficient dynamics learning capability with adequate disturbance rejection characteristics. At the start, a nominal model from a largely unknown DFUAV model is achieved offline through sparse identification. Afterward, a reinforcement learning (RL) mechanism is deployed online to learn a policy to facilitate the initial guesses for the control input sequence. Thereafter, an MPC-driven optimization problem is developed, where the obtained nominal (learned) system is updated by the real system, yielding improved computational efficiency for the overall control framework. Under appropriate assumptions, stability and recursive feasibility are compactly ensured. Finally, a comparative study is conducted to illustrate the efficacy of the designed scheme.

In this article, we consider the "nearest distance" from a given uncontrollable dynamical network to the set of controllable ones. We consider networks whose behaviors are represented via linear dynamical systems. The problem of interest is then finding the smallest number of entries/parameters in the system matrices, corresponding to the smallest number of edges of the networks, that need to be perturbed to achieve controllability. Such a value is called the zero-norm distance to controllability (ZNDC). We show genericity exists in this problem, so that other matrix norms (such as the 2-norm or the Frobenius norm) adopted in this notion are nonsense. For ZNDC, we show it is NP-hard to compute, even when only the state matrices can be perturbed. We then provide some nontrivial lower and upper bounds for it. For its computation, we provide two heuristic algorithms. The first one is by transforming the ZNDC into a problem of structural controllability of linearly parameterized systems, and then greedily selecting the candidate links according to a suitable objective function. The second one is based on the weighted -norm relaxation and the convex-concave procedure, which is tailored for ZNDC when additional structural constraints are involved in the perturbed parameters. Finally, we examine the performance of our proposed algorithms on several typical uncontrollable networks arising in multiagent systems.

In the field of 3D shape recognition, the view-based approach has achieved state-of-the-art performance. A major challenge that needs to be addressed by the view-based approach is how to effectively aggregate multi-view features to obtain a better 3D shape representation. Existing methods which rely on networks with static parameters for feature aggregation adversely coerce the network to learn a general feature aggregation strategy for all inputs, ignoring the diversity of input 3D shapes in real-world scenarios. In this work, we propose a novelDynamic View Aggregation NetworkcalledDVA-Netto address this challenge. DVA-Net can dynamically adjust the network parameter depending on the input 3D shapes to flexibly fuse multi-view information. The shape-specific parameter adaptation is achieved by our designedDynamic Relation-aware Aggregationmodule, dubbedDRAmodule. It is responsible for learning relations among views and adaptively integrating multi-view features. Comprehensive experiments on benchmark datasets demonstrate that our proposed method achieves state-of-the-art performance for 3D shape classification and retrieval.

This article proposes a novel reinforcement learning-based model predictive control (RLMPC) scheme for discrete-time systems. The scheme integrates model predictive control (MPC) and reinforcement learning (RL) through policy iteration (PI), where MPC is a policy generator and the RL technique is employed to evaluate the policy. Then the obtained value function is taken as the terminal cost of MPC, thus improving the generated policy. The advantage of doing so is that it rules out the need for the offline design paradigm of the terminal cost, the auxiliary controller, and the terminal constraint in traditional MPC. Moreover, RLMPC proposed in this article enables a more flexible choice of prediction horizon due to the elimination of the terminal constraint, which has great potential in reducing the computational burden. We provide a rigorous analysis of the convergence, feasibility, and stability properties of RLMPC. Simulation results show that RLMPC achieves nearly the same performance as traditional MPC in the control of linear systems and exhibits superiority over traditional MPC for nonlinear ones.

This article develops a distributed economic model predictive control (EMPC) method which is applied in a group of interconnected linear subsystems subject to unknown bounded disturbances. Multiple subsystems are coupled through the dynamics, and the control objective is to optimize some general performance criteria of the whole system which may take economic considerations into account. First, a two operation modes EMPC optimization problem is formulated, which incorporates the constraints derived from the Lyapunov technique. In the first mode, each subsystem focuses on the optimization of the economic performance while maintaining the state in a certain region. In the second mode, the system states are steered to a neighborhood of a steady state by making use of the Lyapunov-based constraints. Furthermore, a consensus alternating direction method of multipliers (ADMM) is adopted to solve the model predictive control optimization problems with a coupled predicted model constraint in a distributed way. By introducing consensus constraints, the resulting local optimization problem does not depend on real-time optimal solutions from neighboring subsystems and allows subsystems to solve it in parallel. Moreover, the closed-loop system is ensured to be input-to-state stable (ISS) with respect to the disturbances. To demonstrate the effectiveness of the algorithm, we conduct numerical simulations on a thermal power interconnected system.

In this article, we propose a novel economic model-predictive control (MPC) algorithm for a group of disturbed linear systems and implement it in a distributed manner. The system consists of multiple subsystems interacting with each other via dynamics and aims to optimize an economic objective. Each subsystem is subject to constraints both on states and inputs as well as unknown but bounded disturbances. First, we divide the computation of control inputs into several local optimization problems based on each subsystem's local information. This is done by introducing compatibility constraints to confine the difference between the actual information and the previously published reference information of each subsystem, which is the key feature of the proposed distributed algorithm. Then, to ensure the satisfaction of both state and input constraints under disturbances, constraints are tightened on the state and the input of nominal systems by considering explicitly the effect of uncertainties. Moreover, based on an overall optimal steady state, a dissipativity constraint and a terminal constraint are designed and incorporated in the local optimization problems to establish recursive feasibility and guarantee stability for the resulting closed-loop system. Finally, the efficiency of the distributed economic MPC algorithm is demonstrated in a building temperature control case study.

This article studies the formation and trajectory tracking control of multiagent systems. We present a novel multilayer graph for the multiagent system to enable extensibility of the interaction network. Based on the multilayer graph, a formation control law by using the potential function approach is developed for autonomous formation, formation maintenance, collision, and obstacle avoidance. When the desired formation is achieved, the barycentric of the formation shape is viewed as a virtual leader, and a model predictive control (MPC) scheme is applied to the virtual leader for tracking a reference trajectory; meanwhile, the agents will maintain the desired angles and distances via the formation control law. By applying the proposed schemes, the tasks of formation maintenance and trajectory tracking in a constrained space are fulfilled. Comprehensive simulation studies under different environmental constraints and trajectories confirm the effectiveness of the proposed approaches in addressing the formation and trajectory tracking problems.

This article develops a dynamic version of event-triggered model predictive control (MPC) without utilizing any terminal constraint. Such a dynamic event-triggering mechanism takes the advantages of both event- and self-triggering approaches by dealing explicitly with conservatism in the triggering rate and measurement frequency. The prediction horizon shrinks as the system states converge; we prove that the proposed strategy is able to stabilize the system even without any stability-related terminal constraint. Recursive feasibility of the optimization control problem (OCP) is also guaranteed. The simulation results illustrate the effectiveness of the scheme.

In this article, we investigate the fixed-time behavioral control problem for a team of second-order nonlinear agents, aiming to achieve a desired formation with collision/obstacle avoidance. In the proposed approach, the two behaviors(tasks) for each agent are prioritized and integrated via the framework of the null-space-based behavioral projection, leading to a desired merged velocity that guarantees the fixed-time convergence of task errors. To track this desired velocity, we design a fixed-time sliding-mode controller for each agent with state-independent adaptive gains, which provides a fixed-time convergence of the tracking error. The control scheme is implemented in a distributed manner, where each agent only acquires information from its neighbors in the network. Moreover, we adopt an online learning algorithm to improve the robustness of the closed system with respect to uncertainties/disturbances. Finally, simulation results are provided to show the effectiveness of the proposed approach.