Packet-based control: The H2H2-optimal solution

This paper presents a new control method for networked control systems. Motivated by a more efficient use of the packet structure, multipoint packets are used to reduce network traffic and computation time. The control problem for the multipoint-packet system is shown to be equivalent to a multirate control problem, which in turn is reduced to a synthesis problem with a constraint on the feedthrough matrix. This problem is solved in general using H2H2 optimization techniques and the solution to the general problem is applied to the multipoint-packet control problem. The paper concludes with an example that compares the performance of the proposed method to more traditional ones.