Showing papers by "Maria Domenica Di Benedetto published in 2022"

Scalable Mesh Stability of Nonlinear Interconnected Systems

Abstract: This letter deals with large-scale interconnected systems with general network topology, affected by external disturbances, and with the possibility to ensure the overall stability when some sufficient conditions are met by each agent with respect to its neighbors. We introduce the notion of scalable Mesh Stability (sMS), that requires the existence of trajectory bounds that do not depend on the number of subsystems. The immediate consequence is that perturbations originating in a point of the interconnected system do not amplify through it. A numerical example on interconnection of microgrids shows the interest and the effectiveness of the theoretical result.

Symbolic Control Design of an Artificial Pancreas for Type-2 Diabetes

Abstract: Diabetes is a widespread disease characterized by chronic hyperglycemia so that diabetic individuals usually require the administration of exogenous insulin for survival. As a consequence, in the context of the so-called artificial pancreas, many glucose control methods have been presented in the last few years. In this work, we focus on type-2 diabetes and propose a novel model-based glucose control technique based on the use of symbolic models, which are finite approximations of complex dynamical systems. This framework allows taking into account nonlinearities and delays in the dynamics, uncertainties, and input bounds, as well as nonidealities coming from the interaction between physical plant and digital environment. The methodology is extensively validated over a virtual patient model, broadly accepted as a substitute to animal trials in the preclinical testing of closed-loop glucose control strategies. The results show the effectiveness and the robustness of the approach.

Mesoscopic Controller for String Stability of Platoons With Disturbances

Abstract: This article exploits macroscopic information for the control of autonomous vehicles in platoon formation in case of external disturbances. The use of such information leads to smoother platooning. A mesoscopic controller is proposed, and disturbance string stability is proven through input-to-state stability concepts. Simulations prove the efficacy of the proposed approach by showing its robustness with respect to the presence of perturbations acting on the platoon vehicles.

Output Feedback Reachability of Controlled–Observable States for Nondeterministic Finite–State Systems

Abstract: In this letter control design of nondeterministic finite state systems with reachability specifications is addressed. The class of controllers we use is rather general and combines feedforward and output feedback schemes. The proposed controller allows not only the state of the system to reach the desired target set but also the identification of which state of the target set has been reached. Necessary and sufficient conditions are derived for the control problem to admit a solution and a controller is designed. The solution to the investigated problem has important implications in the context of recovery control and symbolic control design of nonlinear and hybrid systems, as discussed also through some examples.

Sampled-data glucose regulation with reach-and-stay specifications through time-varying contracts

Abstract: The complexity of the glucose-insulin system makes the glucose control problem a hard task to accomplish. In this context, a decentralized approach can be of help, through the exploitation of contracts theory, which allows to formalize the fulfillment of safety/invariance specifications over a system in terms of set of assumptions and guarantees over the composing subsystems. We here take a compositional model-based approach considering, as a first attempt, simplified scalar glucose and insulin subsystems. Assumptions and guarantees sets are piecewise-constant time-varying intervals, computed at sampling times, on the basis of the glucose measurements, so they are not completely known a priori. Updating the intervals may lead to temporary violation of the contracts, according to their classical definition, until the system reaches the new target set. By exploiting the property of monotonicity of the involved subsystems, we define a minimum-time reachability problem, which is solved in closed form to minimize the worst-case contract time violation, and such that the insulin subsystem is steered to a controlled invariant set (reach-and-stay specification). Simulations performed in a non-ideal scenario confirm the potential of the proposed approach.

Path generation based on convex lifting: optimization of the corridors

Abstract: The capability to compute a partition of a cluttered environment starting from the obstacles that lie in it enables the construction of a route connecting an initial point to a desired final one. In this paper, we revisit the convex lifting method for the computation of a suitable partition of the cluttered environment with the objective to propose a systematic procedure of reorganization of the cells within the partition in order to repel their boundaries from the obstacles. The ultimate goal is the construction of a connecting-path between an initial point and a final point characterized by a corridor with improved width guaranteeing the collision avoidance. The qualities of the corridors will impact the constraints on the motion of the controlled agent and consequently the real-time performance and robustness of the navigation in a cluttered environment.

Deep reinforcement learning for closed-loop blood glucose control: two approaches

Abstract: Reinforcement learning, thanks to the observation-action approach, represents a useful control tool, in particular when the dynamics are characterized by strong non-linearity and complexity. In this sense, it has a natural application in the biological systems field where the complexity of the dynamics makes the automatic control particularly challenging. This paper presents a combined application of neural networks and reinforcement learning, in the so-called field of deep reinforcement learning, for the glucose regulation problem in patients with diabetes mellitus. The glucose control problem is solved through the Deep Deterministic Policy Gradient (DDPG) and the Soft Actor-Critic (SAC) algorithms, where the environment exploited for the agent's interactions is represented by a glucose model that is completely unknown to agents. Preliminary results show that the DDPG and SAC agents can suitably control the glucose dynamics, making the proposed approach promising for further investigations. The comparison between the two agents shows a better behaviour of SAC algorithm.