Synthesis of a Two Degrees of Freedom Wireless Networked Digital Servo Control System With Dynamic State Prediction Based on a Gated Recurrent Unit-Based Round Trip Time Predictor

A wireless networked control system (WNCS) may suffer from an abrupt and large network communication delay that may degrade the control performance and lead to the instability of the closed-loop system. A two degrees of freedom (DOF) digital servo system for a WNCS is proposed in this paper. We trained a gated recurrent unit network to predict the round trip time (RTT) of a closed-loop packet with high accuracy and integrated it with a feedback controller connected to the network. The feedback controller was configured to monitor the behavior of RTTs for each node in the network, predict the future behavior of the nodes, and synthesize a control input that compensates for the worst case RTT of the nodes using the dynamic state predictive control scheme. We constructed a pendulum manipulator system, in which an inverted pendulum is attached at the tip of the six DOF industrial manipulator. The control objective of a pendulum manipulator system was to track the first axis angle of the manipulator to the given reference trajectory while stabilizing the pendulum around its unstable equilibrium. The results of the numerical simulations and experiments were reported to validate the performance of the proposed controller in a WNCS that occasionally suffers from a large network transmission delay that amounted to several tens of sampling intervals at the worst case.