The event-triggered sampling optimization criterion for distributed networked monitoring and control systems

An overview of signal-dependent sampling methods in control systems with an emphasis on the event-triggered schemes is done. The analysis of either the networked control system performance, or the sampling cost component measures is performed. A new criterion for sampling a continuous-time variable using the event-triggered pattern is presented. The suggested criterion is based on the optimization of a balance between the signal tracking performance and the cost of a networked control system. As the system cost we consider the amount of communication bandwidth dedicated to sending messages and the length of the FIFO output buffer. The evaluation of particular cost components is made on the basis of timing parameters of a sample string, i.e. the average sampling rate and sample string burstiness. The principles of selection of the sampling interval bounds are discussed. The rules of tuning of the objective function coefficients are presented.