The paper concerns the Linear Quadratic non-Gaussian (LQnG) sub-optimal control problem when the inputand output signals travel through an unreliable network,namely Gilbert-Elliot channels. In particular, the input/outputpacket losses are modeled by Bernoulli sequences, and weassume that the moments of the non-Gaussian noises up tothe fourth order are known. By mean of a suitable rewritingof the system through an intermittent output injection term,and by considering an augmented system with the second-orderKronecker power of the measurements, a simple solution is pro-vided by substituting the Kalman predictor with intermittentobservations of the LQG control law with a quadratic optimalpredictor. Numerical simulations show the effectiveness of theproposed method.
LQ non-Gaussian Control with I/O packet losses
CACACE F;
2020-01-01
Abstract
The paper concerns the Linear Quadratic non-Gaussian (LQnG) sub-optimal control problem when the inputand output signals travel through an unreliable network,namely Gilbert-Elliot channels. In particular, the input/outputpacket losses are modeled by Bernoulli sequences, and weassume that the moments of the non-Gaussian noises up tothe fourth order are known. By mean of a suitable rewritingof the system through an intermittent output injection term,and by considering an augmented system with the second-orderKronecker power of the measurements, a simple solution is pro-vided by substituting the Kalman predictor with intermittentobservations of the LQG control law with a quadratic optimalpredictor. Numerical simulations show the effectiveness of theproposed method.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
