Distributed DSP processing for multivariable state equations

The paper deals with the synthesis of a multivariable digital controller when its hardware must be decentralized. Thus, it can be used as a way to implement dynamic equations in state space for hardware with limited inputs and outputs. The method consists in breaking equations, while keeping the overall transfer matrix, resulting in a distributed implementation. The solution has a modular structure for an arbitrary number of inputs and outputs. A way of performing the synthesis is shown by using the same dynamic matrix. As a consequence, the computational demand could not be severely reduced when compared with a centralized controller. Nevertheless, the synthesis is actually distributed, since each part is implemented with a smaller number of inputs and outputs. A multivariable controller of an inverted pendulum is used as an example.