Robust suboptimal control over lossy networks using extended input schemes

We consider an optimal control problem for networked, linear systems, where the loop is closed via a lossy network with a lossy acknowledgment mechanism. So, the data send from the controller to the actuators might be lost. In these cases often specific strategies called input scheme are used to choose backup inputs. In this work we employ an extended class of such strategies. We derive the necessary optimality conditions for linear controllers. Moreover, we present heuristic methods as well as bilinear matrix inequalities to determine these controllers in a suboptimal way. Examples illustrate that these extended input schemes can improve the performance.