11072018 1531294679

In feedback control systems, a controller determines the input to a dynamical system based on measurements of its outputs. Nowadays, the controller is increasingly often connected to the sensors and actuators of the system via a shared communication network, for example via WiFi. This leads to more flexible systems with easier installation and lower costs. Moreover, it opens up the way to many new control applications, such as cooperative adaptive cruise control, a cruise control system that makes use of wireless communication between vehicles in order to improve fuel efficiency, safety, and traffic throughput. However, as multiple devices share the communication network, and the available bandwidth is often limited, it is important to use the communication network in a careful manner in order not to congest it. Indeed, overuse of the network leads to delays and loss of information, which deteriorate the performance of the system. Hence, it is important to only transmit (e.g., sensor and actuation) data when this is really necessary in order to guarantee the required performance of the system. This can be accomplished by using event-triggered control (ETC).

ETC brings feedback into the communication process by determining the transmission times on-line, based on real-time knowledge of the system. This renders ETC well-equipped to balance control performance and network utilization. However, this also means that transmission times in ETC systems are influenced by disturbances, which is not the case when using standard periodic communication. Hence, just as the control performance of a system should be robust to disturbances, also the communication properties should be robust in the sense that disturbances should not have a large adverse effect on the (minimum and average) inter-transmission times.

Download Thesis