Conservation of bandwidth for networked control systems under time delays and packet dropouts using hidden Markov models

Abstract This paper introduces a new method for enhancing bandwidth conservation in networked control systems (NCSs) through the integration of the single model scheme (SMS) with event-triggered predictive control (ETPC). NCSs face significant challenges from network-induced imperfections, such as time delays and packet dropouts which can compromise system stability and performance. The existing ETPC schemes assume a known bound on the time delays and packet dropouts and send futuristic input commands based on that bound accounting for the worst-case scenario, which overloads the network creating further congestion through bandwidth consumption. The proposed SMS-ETPC scheme addresses these limitations by employing a model that predicts both time delays and packet dropouts using a semi-continuous hidden Markov model (SCHMM) enhanced with Dirac-delta functions to account for the discrete nature of packet dropouts. Additionally, by incorporating event-triggered mechanisms (ETMs) in both feedback and forward channels, the SMS-ETPC framework reduces unnecessary data transmissions, conserves bandwidth consumption and accounts for time delays and packet dropouts. Simulation results demonstrate the efficiency of the SMS-ETPC scheme due to its inherent advantage of conservation of bandwidth in NCSs.