Earliest-arrival path: A global optimized transmission for networked control systems

Abstract

According to existed switched system theorem, packet-loss rate determines the stability of NCSs. With the core purpose to provide low packet-loss rate for distributed networked control systems (NCSs) under constraint communication environment, the transmission optimization problem is discussed from global communication aspect. Firstly, network induced delay is described as time-correlative function by auto-regressive-moving-average (ARMA) model, and the future delay is predicted by the obtained delay samples. Secondly, the widely cited shortest-path-of-time-dependent-network (SPTDN) algorithm is discussed and improved through inserting waiting-time before packets are sent to next node in a path, in order to obtain earlier arrival time than SPTDN. Then the notion of earliest-arrival time (EAT) is proposed to measure the quality of communication path, the online EAT algorithm is designed to search for future EAT and the corresponding path from the global network dynamically. Benefit from the dynamic EAT path, packets arrive at destination as early as possible, it decreases packet-loss rate, hence EAT path ensures the stability of distributed NCSs from the view of global network. Finally, the experiment proves the validity and effectiveness of the proposed approach.