Special Issue: Adaptive Methods for Resilient Control Systems

Resilience literally is the ability to readily recover from illness, depression, or adversity. Control system resilience system's accommodate its faults, disturbances, and uncertainties, desired performance. Resilient control increase for performance guarantees in presence of uncertain faults (such as actuator sensor failures structure damage) disturbances. Desired includes asymptotic tracking optimality, addition stability, their robustness. systems need which needs be provided by effective techniques. Thus, a main feature resilient that they can crucial needed performance-critical applications, design. Research has been developing fast recent years, reflected increased activities either general conferences special conferences, resulted publications. New design methods techniques are emerging applications. Contributing foundations techniques, fault detection, fault-tolerant made significant progresses, so does adaptive unique capabilities parametric, structural, environmental uncertainties caused payload variations aging, component failures, external For systems, face new challenges: how effectively estimate compensate Systematic study handling an open area research with focus guarantee performance, most important control. The goal this issue show state-of-the-art developments advanced using deal recovery This contain six papers cover both theory, applications systems. paper “Robust nonlinear adaptation algorithms multi-task prediction networks” Abulikemu Abuduweili Changliu Liu studies robust online adapt models, lead more accurate predictions facilitate subsequent decisions. There multiple ways results enter loop make resilient. In safety critical situation, reduce uncertainty model, hence reducing margin safe law. During human-robot collaboration, improve accuracy human intention prediction, allowing robot better collaborate choosing actions respond human's needs. Nonetheless, focused solely on adaptable prediction. authors introduced algorithm MEKF-MA-ME combining exponential moving average multi-epoch update modified EKF algorithm. proposed complex models encoded neural networks, validated through numerical studies. second “An decoupling approach unmeasurable coupled dynamics” K. Merve Dogan, Tansel Yucelen, Jonathan A. Muse. Adaptive have capability provide resiliency against operational anomalies such disturbances uncertainties. However, stability challenged dynamics in, example, airplanes high aspect ratio wings, slung load cooperative manipulators, actuated-unactuated dynamical proposes model reference algorithms. key framework it convergence between trajectories given without relying any measurements under tight sufficient condition. also provides generalization address effectiveness matrix, where resulting condition case relies linear matrix inequalities. “Diagnosis assessment secondary protection 68-bus power system” Mustafa Salman, N. Eva Wu, Qiu Qin concerns diagnosis recently developed (SP) applied IEEE test subject primary (PP) misoperations transmission short circuit faults. A PP may fail trip falsely line. aim SP network providing timely corrective action before misoperation causes instability. Using hybrid simulations driven randomly generated discrete events governing rotor angles 16 interconnected synchronous generators, multiple-model filter-based successfully recovered 98 processing time less than 150 ms. “Adaptive compensation persistent systems” Liyan Wen, Gang Tao, Bin Jiang, designs functions characterizations (i) some unknown inputs stuck at fixed varying values instants, (ii) failure pattern always switches one another switching not stop. Such controlled plant described time-varying system, feedback linearization based studied closed-loop stabilization bounded output specific conditions. To scheme developed, indicator function contains formulation. All above certain is, recovering desirable failures. fifth “Mission independent heterogeneous multiagent virtual actuator” Meysam Yadegar Nader Meskin. Fault tolerant (FTC) approaches paper, FTC multi-agent (MAS) loss additive well disturbance investigated. objective keep after occurrence similar healthy one. envisaged each agent require separate isolation, identification unit nor information regarding mission MAS. Finally, “Reinforcement learning flight design” Burak Yuksek Gokhan Inalhan, “resilience adaptation” designed, novel reinforcement (RL)-based (CRM) system. further enhances transient response beyond existing (MRAC) classical CRM-adaptive wide range parametric law implements RL, actor-critic learn policy observer-like gain error observations environment. It shown damped improving oscillatory behavior result instabilities catastrophic accidents if considered. We would like thank interest technical contributions, referees effort reviewing submissions. journal's Managing Editor Professor Michael Grimble, editorial board, office support help issue.