A systematic design methodology for state observers for a large class of nonlinear systems with bounded exogenous inputs (disturbance inputs and sensor noise) is proposed. The nonlinearities under consideration are characterized by an incremental quadratic constraint parameterized by a set of multiplier matrices. Linear matrix inequalities are developed to construct observer gains, which ensure that a performance output based on the state estimation error satisfies a prescribed degree of accuracy. Furthermore, conditions guaranteeing estimation of the unknown inputs to arbitrary degrees of accuracy are provided. The proposed scheme is illustrated with a numerical example, which does not satisfy the so-called “matching conditions.”

Objective: The development of artificial pancreas (AP) technology for deployment in low-energy, embedded devices is contingent upon selecting an efficient control algorithm for regulating glucose in people with type 1 diabetes mellitus. In this paper, we aim to lower the energy consumption of the AP by reducing controller updates, that is, the number of times the decision-making algorithm is invoked to compute an appropriate insulin dose. Methods: Physiological insights into glucose management are leveraged to design an event-triggered model predictive controller (MPC) that operates efficiently, without compromising patient safety. The proposed event-triggered MPC is deployed on a wearable platform. Its robustness to latent hypoglycemia, model mismatch, and meal misinformation is tested, with and without meal announcement, on the full version of the US-FDA accepted UVA/Padova metabolic simulator. Results: The event-based controller remains on for 18 h of 41 h in closed loop with unannounced meals, while maintaining glucose in 70-180 mg/dL for 25 h, compared to 27 h for a standard MPC controller. With meal announcement, the time in 70-180 mg/dL is almost identical, with the controller operating a mere 25.88% of the time in comparison with a standard MPC. Conclusion: A novel control architecture for AP systems enables safe glycemic regulation with reduced processor computations. Significance: Our proposed framework integrated seamlessly with a wide variety of popular MPC variants reported in AP research, customizes tradeoff between glycemic regulation and efficacy according to prior design specifications, and eliminates judicious prior selection of controller sampling times.

This article presents a data-driven method for computing reachable sets where active learning (AL) is used to reduce the computational burden. Set-based methods used to estimate reachable sets typically do not scale well with the state-space dimension, or rely heavily on the existence of a model. If such a model is not available, it is simple to generate state trajectory data by numerically simulating black-box oracles of systems (whose dynamics are unknown) from sampled initial conditions. Using these data samples, the estimation of reachable sets can be posed as a classification problem, wherein AL can intelligently select samples that are most informative and least similar to previously labeled samples. By exploiting submodularity, the actively learned samples can be selected efficiently, with bounded suboptimality. Our proposed framework is illustrated by estimating the domains of attractions of model predictive controllers (MPCs) and reinforcement learners. We also consider a scenario where there are two oracles that differ with respect to evaluation costs and labeling accuracy. We propose a framework to reduce the dependency of the expensive oracle in labeling samples using disagreement-based AL (DBAL). The potential of the DBAL algorithm is demonstrated on a solver selection problem for real-time MPC.

This article addresses the problem of state estimation and simultaneous learning of the vehicle's tire model on autonomous vehicles. The problem is motivated by the fact that lateral distance measurements are typically available on modern vehicles while tire models are difficult to identify and also vary with time. Tire forces are modeled in the estimator using a neural network in which no a priori assumptions on the type of model need to be made. A neuro-adaptive observer that provides asymptotically stable estimation of the state vector and of the neural network weights is developed. The developed observer is evaluated using both MATLAB simulations with a low-order model as well as with an unknown high-order model in the commercial software CarSim. Cornering and lane change maneuvers are used to learn the tire model over an adequately large range of slip angles. Performance with the low-order vehicle model is excellent with near-perfect estimation of states as well as the tire force nonlinear characteristics. The performance with the unknown high-order CarSim model is also found to be good with the tire model being estimated correctly over the range of slip angles excited by the executed vehicle maneuvers. The developed technology can enable a new approach to obtaining tire models that are otherwise difficult to identify in practice and depend on empirical characterizations.

Wireless networked control systems (WNCSs) have the potential to revolutionize industrial automation in smart factories. Optimizing closed-loop performance while maintaining stability is a fundamental challenge in WNCS due to limited bandwidth and nondeterministic link quality of wireless networks. In order to bridge the gap between network design and control system performance, we propose an optimal dynamic transmission scheduling strategy that optimizes the performance of multiloop control systems by allocating network resources based on predictions of both link quality and control performance at run time. We formulate the optimal dynamic scheduling problem as a nonlinear integer programming problem, which is relaxed to a linear programming problem. We further extend the optimization problem to balance control performance and communication cost. The proposed optimal dynamic scheduling strategy renders the closed-loop system mean-square stable under mild assumptions. Its efficacy is demonstrated by simulating a four-loop control system over an IEEE 802.15.4 wireless network simulator--TOSSIM. The run-time network reconfiguration protocol tailored for optimal scheduling is designed and implemented on a real wireless network consisting of IEEE 802.15.4 devices. Hybrid simulations integrating a real wireless network and simulated physical plant control are performed. Simulation and experimental results show that the optimal dynamic scheduling can enhance control system performance and adapt to both constant and variable wireless interference and physical disturbance to the plant.