Output feedback MPC with send-on-delta measurements for uncertain systems

In many applications, it is desired to reduce the communications between sensors and the controller for example to reduce the network load or to improve battery lifetime. We present an output feedback model predictive control approach for uncertain systems subject to measurement noise that utilizes the event driven-so called send-on-delta strategy Based on the change of the measurement, the sensor decides, if it sends data to the controller. The controller combines a linear state estimator with a predictive control law. It employs an online approach to improve the error bounds and thus the control performance. Conditions that guarantee closed loop properties are presented. We illustrate the results by an example.