A filtered predictive PI controller for wirelessHART networked control system

In recent years, increasing attention has been paid towards applying wireless technology for control. This is due to its advantages of exibility, scalability, use of fewer cables and overall reduced operational cost compared to its wired counterpart. However, the technology is often affected by stochastic delay and high frequency noise. PIDs are ill-equipped to deal with these problems while model based controllers such as dead-time compensators (DTCs) like Smith predictor and internal model controllers (IMCs) are complex and require exact plant model for implementation. Thus, predictive PI (PPI) controller being a settlement between the PIDs and the model based controllers is a good candidate. The PPI retains simplicity of the PID, it has the ability to predict long time delay and can be used even with model mismatch. However, the PPI is severely affected by high frequency noise. Therefore, this paper proposes a Filtered PPI controller that can be used even in the presence of high frequency noise. Simulation and experimental results proved the viability of the proposed method.