Remanufacturing is fundamentally more challenging than traditional manufacturing due to the significant uncertainty, variability, and incompleteness inherent in end-of-life (EoL) products. At the same time, it has become increasingly essential and urgent for facilitating a circular economy, driven by the growing volume of discarded electronic products and the escalating scarcity of critical materials. In this paper, we review the existing literature and examine the key challenges as well as emerging opportunities in intelligent automation for EoL electronics remanufacturing, providing a comprehensive overview of how robotics, control, and artificial intelligence (AI) can jointly enable scalable, safe, and intelligent remanufacturing systems. This paper starts with the definition, scope, and motivation of remanufacturing within the context of a circular economy, highlighting its societal and environmental significance. Then it delves into intelligent automation approaches for disassembly, inspection, sorting, and component reprocessing in this domain, covering advanced methods for multimodal perception, decision-making under uncertainty, flexible planning algorithms, and force-aware manipulation. The paper further reviews several emerging techniques, including large foundation models, human-in-the-loop integration, and digital twins that have the potential to support future research in this area. By integrating these topics, we aim to illustrate how next-generation remanufacturing systems can achieve robust, adaptable, and efficient operation in the face of complex real-world challenges.
Mean field games (MFGs) offer a powerful framework for modeling large-scale multi-agent systems. This paper addresses MFGs formulated in continuous time with discrete state spaces, where agents' dynamics are governed by continuous-time Markov chains – relevant to applications like population dynamics and queueing networks. While prior research has largely focused on theoretical aspects of continuous-time discrete-state MFGs, efficient computational methods for determining equilibria remain underdeveloped. Inspired by discrete-time approaches, we approximate the classical Nash equilibria by regularization methods, enabling more computationally tractable solution algorithms. Specifically, we define regularized equilibria for continuous-time MFGs and extend the classical fixed-point iteration and fictitious play algorithm to these equilibria. We validate the effectiveness and practicality of our approach via illustrative numerical examples.
This paper develops certificates that propagate compatibility of multiple control barrier function (CBF) constraints from sampled vertices to their convex hull. Under mild concavity and affinity assumptions, we present three sufficient feasibility conditions under which feasible inputs over the convex hull can be obtained per coordinate, with a common input, or via convex blending. We also describe the associated computational methods, based on interval intersections or an offline linear program (LP). Beyond certifying compatibility, we give conditions under which the quadratic-program (QP) safety filter is affine in the state. This enables explicit implementations via convex combinations of vertex-feasible inputs. Case studies illustrate the results.
This work investigates a two-player coordination game in which the players exhibit heterogeneous levels of bounded rationality. We analyze the log-linear learning dynamics where the probability distribution used to select which of the agents gets to revise its strategy is fixed but not necessarily uniform. The stationary distribution of the resulting Markov chain on the strategy profile space is derived in closed-form as a function of the rationalities and the agent selection probabilities. We proceed by showing that adjusting the selection probabilities can be used to bias the stationary distribution toward the potential-maximizing state. However, this optimization comes at the cost of a reduced convergence rate, whereas the uniform selection probabilities uniquely maximizes the convergence speed irrespective of the players’ rationality levels. A Pareto-optimal probability selection rule is proposed, trading-off the distributional bias with convergence rate. Moreover, it is shown that in coordination games, high levels of rationality sometimes accelerate convergence, whereas in other cases they may paradoxically hinder the convergence rate of log-linear learning dynamics.
This paper proposes a robust tube-based model predictive control approach for spacecraft rendezvous subject to sector-bounded nonlinearities and bounded disturbances. Unlike existing methods that design the feedback controller and tube tightening parameters sequentially, we jointly optimize both through a convex Linear Matrix Inequality using static quadratic constraints. This eliminates the conservatism inherent in two-step design procedures, while maintaining computational tractability for real-time implementation. The approach is validated through a CubeSat docking simulation with tight operational constraints, showing around 70% fuel savings, 21% reduction in average computational time and smaller tube sizes compared to LQR-based fixed-gain methods.
Robust adaptive control methods are essential for maintaining quadcopter performance under external disturbances and model uncertainties. However, fragmented evaluations across tasks, simulators, and implementations hinder systematic comparison of these methods. This paper introduces an easy-to-deploy, modular simulation testbed for quadcopter control, built on RotorPy, that enables evaluation under a wide range of disturbances such as wind, payload shifts, rotor faults, and control latency. The framework includes a library of representative adaptive and non-adaptive controllers and provides task-relevant metrics to assess tracking accuracy and robustness. The unified modular environment enables reproducible evaluation across control methods and eliminates redundant reimplementation of components such as disturbance models, trajectory generators, and analysis tools. We illustrate the testbed's versatility through examples spanning multiple disturbance scenarios and trajectory types, including automated stress testing, to demonstrate its utility for systematic analysis. Code is available at https://github.com/Dz298/AdaptiveQuadBench.
Rapid robot motion generation is critical in Human-Robot Collaboration (HRC) systems, as robots need to respond to dynamic environments in real time by continuously observing their surroundings and replanning their motions to ensure both safe interactions and efficient task execution. Current sampling-based motion planners face challenges in scaling to high-dimensional configuration spaces and often require post-processing to interpolate and smooth the generated paths, resulting in time inefficiency in complex environments. Optimization-based planners, on the other hand, can incorporate multiple constraints and generate smooth trajectories directly, making them potentially more time-efficient. However, optimization-based planners are sensitive to initialization and may get stuck in local minima. In this work, we present a novel learning-based method that utilizes a Flow Matching model conditioned on a single-view point cloud to learn near-optimal solutions for optimization initialization. Our method does not require prior knowledge of the environment, such as obstacle locations and geometries, and can generate feasible trajectories directly from single-view depth camera input. Simulation studies on a UR5e robotic manipulator in cluttered workspaces demonstrate that the proposed generative initializer achieves a high success rate on its own, significantly improves the success rate of trajectory optimization compared with traditional and learning-based benchmark initializers, requires fewer optimization iterations, and exhibits strong generalization to unseen environments.
In Bayesian persuasion, an informed sender, who observes a state, commits to a randomized signaling scheme that guides a self-interested receiver's actions. Classical models assume the receiver knows the commitment. We, instead, study the setting where the receiver infers the scheme from repeated interactions. We bound the sender's performance loss relative to the known-commitment case by a term that grows with the signal space size and shrinks as the receiver's optimal actions become more distinct. We then lower bound the samples required for the sender to approximately achieve their known-commitment performance in the inference setting. We show that the sender requires more samples in persuasion compared to the leader in a Stackelberg game, which includes commitment but lacks signaling. Motivated by these bounds, we propose two methods for designing inferable signaling schemes, one being stochastic gradient descent (SGD) on the sender's inference-setting utility, and the other being optimization with a boundedly-rational receiver model. SGD performs best in low-interaction regimes, but modeling the receiver as boundedly-rational and tuning the rationality constant still provides a flexible method for designing inferable schemes. Finally, we apply SGD to a safety alert example and show it to find schemes that have fewer signals and make citizens' optimal actions more distinct compared to the known-commitment case.
Numerous elements drive the spread of infectious diseases in complex real-world networks. Of particular interest is social behaviors that evolve in tandem with the spread of disease. Moreover, recent studies highlight the importance of understanding how multiple strains spread simultaneously through a population (e.g. Delta and Omicron variants of SARS-CoV-2). In this paper, we propose a bi-virus SIS epidemic model coupled with a game-theoretic social distancing behavior model. The behaviors are governed by replicator equations from evolutionary game theory. The prevalence of each strain impacts the choice of an individual to social distance, and, in turn, their behavior affects the spread of each virus in the SIS model. Our analysis identifies equilibria of the system and their local stability properties, which reveal several isolated fixed points with varying levels of social distancing. We find that endemic co-existence is possible only when the reproduction numbers of both strains are equal. Assuming the reproduction number for each virus is the same, we identify suitable parameter regimes that give rise to lines of coexistence equilibria. Moreover, we also identify conditions for local exponential stability of said lines of equilibria. We illustrate our findings with several numerical simulations.
Robust control seeks stabilizing policies that perform reliably under adversarial disturbances, with ℋ_∞ control as a classical formulation. It is known that policy optimization of robust ℋ_∞ control naturally lead to nonsmooth and nonconvex problems. This paper builds on recent advances in nonsmooth optimization to analyze discrete-time static output-feedback ℋ_∞ control. We show that the ℋ_∞ cost is weakly convex over any convex subset of a sublevel set. This structural property allows us to establish the first non-asymptotic deterministic convergence rate for the subgradient method under suitable assumptions. In addition, we prove a weak Polyak-Łojasiewicz (PL) inequality in the state-feedback case, implying that all stationary points are globally optimal. We finally present a few numerical examples to validate the theoretical results.
We consider federated learning of linearly-parameterized nonlinear systems. We establish theoretical guarantees on the effectiveness of federated nonlinear system identification compared to centralized approaches, demonstrating that the convergence rate improves as the number of clients increases. Although the convergence rates in the linear and nonlinear cases differ only by a constant, this constant depends on the feature map ϕ, which can be carefully chosen in the nonlinear setting to increase excitation and improve performance. We experimentally validate our theory in physical settings where client devices are driven by i.i.d. control inputs and control policies exhibiting i.i.d. random perturbations, ensuring non-active exploration. Experiments use trajectories from nonlinear dynamical systems characterized by real-analytic feature functions, including polynomial and trigonometric components, representative of physical systems including pendulum and quadrotor dynamics. We analyze the convergence behavior of the proposed method under varying noise levels and data distributions. Results show that federated learning consistently improves convergence of any individual client as the number of participating clients increases.
We present a conceptual framework for an autonomous safety mechanism designed to enhance the reliability of Unmanned Aerial Vehicles (UAVs) that use Visual-Inertial Odometry (VIO) for state estimation. As UAVs increasingly interact with the public, such safety mechanisms are crucial to reducing the likelihood and severity of accidents. VIO drift, which occurs when accumulated estimation errors cause discrepancies between the UAV’s perceived and actual position, poses a significant risk to safe operation. To address this challenge, we propose a Kalman filter-based approach for detecting VIO drift events. Upon detection, the envisioned safety mechanism is designed to adjust state estimation by integrating onboard gyroscope measurements and thrust commands for short durations, aiming to enhance stability and prevent potential crashes before initiating a controlled landing. While this framework provides the foundation for a real-time safety mechanism, the implementation and experiment focus on validating the drift detection component in an offline setting using real UAV flight data. The results demonstrate the effectiveness of the detection method in identifying VIO drift scenarios, highlighting its potential for future real-time applications.
This paper analyzes the noise sensitivity of the semidefinite program (SDP) used in the direct data-driven infinite horizon linear quadratic regulator (LQR) problem for discrete-time linear time-invariant systems, using a scalar system. While this SDP is shown to find the true LQR controller in the noise-free setting, we show that it leads to a trivial solution when data are corrupted by noise, even when the noise is arbitrarily small. Hence, a "certainty equivalence" approach that uses the original SDP with noisy data is not appropriate.
We obtain conditions for the Wasserstein regularity of non-linear filters, which serve as reduced MDP models for Partially Observable Markov Decision Processes (POMDPs). In particular, we establish a new set of assumptions: instead of requiring total variation continuity of the transition kernel as in earlier work, we show that Wasserstein continuity of the transition kernel, together with total variation continuity of the observation channel, is sufficient to ensure key results. These conditions lead to: (i) unique ergodicity and (ii) geometric ergodicity of the control-free non-linear filter kernel, as well as (iii) existence of optimal policies and (iv) their rigorous approximations and Q-learning methods for both discounted and average cost optimal stochastic control problems. We further identify complementary conditions under which finite-window policies remain near-optimal and provide refined error bounds. Our framework unifies and extends prior results while offering verifiable conditions for analysis and design in POMDPs.
This paper presents a framework supporting enforcement of functional constraints for a range of computational neural models relating to motor control. It builds upon a control structure addressing the tracking problem for a class of neuro-musculoskeletal systems. We illustrate, through a specific example, the manner in which it may be used to investigate functional efficacy of a class of spinal pathway models. Specifically, we assess efficacy of a stretch-reflex model in fulfilling its expected function and compare it to performance to an alternative model proposed here. A discussion of the manner in which the framework can be employed to help constrain a broad range of neural models concludes this paper.
This work analyzes and develops some fundamental results for attitude consensus control of a network of rigid-body vehicles, considered a multi-agent rigid body system (MARBS). The system is analyzed using a full rigid body dynamics model on TSO(3) for each vehicle (agent) in the network. Therefore, the state space of the system is TSO(3)N, where N is the number of vehicles. Attitude synchronization control laws for each vehicle to reach a consensus attitude with zero angular velocity for a particular type of network are obtained, using a Morse-Lyapunov function. Some fundamental results on equilibria of the network under these attitude consensus control laws are obtained. We show that unlike cooperative control of multi-agent systems with highly simplified dynamics models for agents, like point particles or unicycles where the state space of the dynamics is modeled as a vector space, there are multiple equilibrium solutions possible for attitude consensus control laws for a MARBS with dynamics on TSO(3)N. Further, the number of equilibria depends on the network graph topology. This is followed by numerical simulation results for two different network graphs, which show this network control framework to be effective in obtaining attitude consensus.
This paper investigates and compares the convergence performance of two widely used Nonlinear Programming (NLP) solvers-MATLAB's fmincon and IPOPT- for solving multi-impulse cislunar trajectory optimization problems. The problem is formulated as a minimum-fuel (or minimum-triangle upsilon) trajectory optimization for a transfer between two Distant Retrograde Orbits (DROs) in the circular restricted threebody problem. Derivatives of the objective and constraints, required by the solvers, are computed using three methods: a) solver's built-in finite-difference (FD) method, b) the complexstep based (CX) method, and c) an analytical method. The results demonstrate that while analytical derivatives provide the fastest convergence, the CX method achieves nearly the same performance, despite being significantly easier to compute than the analytical derivatives. The CX method outperforms the FD method in terms of derivative accuracy and its impact on the convergence performance of the solvers (i.e., total number of iterations and function evaluations). Results also indicate that IPOPT exhibits faster convergence compared to fmincon, when CX and analytical derivatives are used.
Roll-to-roll (R2R) mechanical transfer is an advanced technology that facilitates continuous, etchant-free transfer of printed electronics and two-dimensional (2D) materials on flexible substrates. It involves dry peeling of fabricated devices or material from a donor substrate and transferring them to a target substrate, both being flexible and known as webs. Achieving high-quality transfer requires precise regulation of the tensions in these webs, necessitating a controller that handles process nonlinearities, disturbances, and input constraints. To address these challenges, we propose a nonlinear model predictive control (MPC) approach for R2R mechanical peeling. However, current R2R peeling models are either too computationally expensive to use or lack the accuracy needed for real-time application. This study introduces a fast, accurate nonlinear model of R2R peeling and integrates it into an MPC framework. A case study on the dry transfer of chemical vapor deposition (CVD) graphene is used to demonstrate the effectiveness of this proposed nonlinear MPC approach.
This study investigates the impact of delays and network structure on the spread dynamics of the Susceptible-Infected-Susceptible (SIS) model using Physics-Informed Neural Networks (PINNs). By integrating network-based models with PINNs, we aim to improve the accuracy of predictions, particularly in the context of delayed and non-delayed transmission scenarios. We demonstrate that the network-based approach yields more precise estimates of spread rates and recovery patterns compared to single-population models. Moreover, the results emphasize the importance of incorporating delays in the estimation and prediction of spread patterns, highlighting the potential of PINNs to enhance our understanding and control of various phenomena in complex networks.
Inner-outer-loop control is widely used for controlling mechanical systems with time-scale separation. Model Predictive Control (MPC) is a popular technique for systems that require command following with state and control constraints. We present a conversion algorithm for MPC-based inner- and outer-loop control by accounting for timing intricacies. For uncertain systems, we apply inner-outer-loop control based on predictive cost adaptive control (PCAC) to flight-control examples, with and without the conversion algorithm. Numerical examples show that inner-outer-loop PCAC improves command following and constraint satisfaction when the conversion algorithm is used. The investigation in this paper is numerical, and thus the contribution is technological rather than theoretical.
