Asynchronous Embedded Model Control for Networked Control Systems

The aim of this thesis is to exploit the Embedded Model Control (EMC) in order to deal with Networked Control Systems (NCS), whose peculiarity is the usage of a communication network (CN) to connect the plant to the control unit, leading to asynchronous measurements and commands. In these types of systems the execution of the control law, implemented remotely, is driven by the arrival of the measures, whose frequency is non-deterministic and variable over time, implying that the controller must deal with variable sampling times. Thus the canonical EMC design needs to be extended in order to face the problems provided by introducing a CN in a closed-loop control system, such extension assumes the name of Asynchronous-EMC (AEMC). More specifically, in this thesis, the effectiveness of the proposed control methodology is tested by applying it to a remote control of a two-wheeled differential drive mobile robot, endowed with a Raspberry Pi board. These tests require the construction of a CN between the robot and a laptop (which will act as a server). For this purpose two types of communication will be tested: the first is based on a p2p layout, built through an Ethernet cable connection, while the other uses a router for forwarding packets from the robot to the server and vice versa. Both of them are tested in order to provide more than one benchmark for the AEMC.