How To Tame Your Sparsity Constraints
We show that designing sparse $H_\infty$ controllers, in a discrete (LTI) setting, is easy when the controller is assumed to be an FIR filter. In this case, the problem reduces to a static output feedback problem with equality constraints. We show how to obtain an initial guess, for the controller, and then provide a simple algorithm that alternates between two (convex) feasibility programs until converging, when the problem is feasible, to a suboptimal $H_\infty$ controller that is automatically stable. As FIR filters contain the information of their impulse response in their coefficients, it is easy to see that our results provide a path of least resistance to designing sparse robust controllers for continuous-time plants, via system identification methods.