Anytime Control Under Practical Communication Models

We investigate a novel anytime algorithm for wireless networked control with random dropouts. The controller computes sequences of tentative future control commands using time-varying (Markovian) computation resources. The sensor-controller and controller-actuator channel states are spatial- and time-correlated and are modeled as a multi-state Markov process. We develop a novel cycle-cost-based approach to obtain stability conditions on the nonlinear plant, controller, network and computation resources.