Event-based Resource Scheduling for Nonlinear Networked Systems Problem description:

Networked Control Systems (NCS) are distributed systems composed of multiple plants, sensors, and controllers exchanging sensory through a shared communication channel. Local systems are often physically isolated, and are solely coupled by the shared communication resource. Each local system tries to send its sensory data to controller utilizing the shared channel. Due to energy and bandwidth limitations however, the communication channel is often subject to some constraints, e.g. a limited number of local systems can simultaneously utilize the channel to update their own controllers. As a conclusion, several subsystems which have not awarded the transmission chance must wait for the channel to become free and then transmit. This might lead to unexpected issues in an NCS such as instability or large errors, especially if the plants are nonlinear. One solution is to design a network manager (scheduler) which decides how to assign the limited communication resource to the local systems such that the stability of the overall NCS is not compromised. The goal of this thesis is designing an error-dependent scheduling law that ensures the stability of the overall NCS. We achieve the goal through linearization of the local systems around their operational points and then design the scheduling law for the linearized systems. We assume that each local nonlinear system is controlled by a local stabilizing controller, i.e. ignoring the communication network, the local systems are stable. Unlike time-triggered policies which assign the communication resource according to a pre-defined ordering, event-based schemes determine the access upon the occurrence of an event. In this thesis we study the stability and performance of an NCS comprised of multiple number of nonlinear plants by designing an event-based scheduler. The scheduler is designed to efficiently assign the communication bandwidth to the subsystems such that the overall NCS is stabilized. Tasks: • Linearization of the nonlinear local plants around the operational point