Collaborative Processing in Distributed Control for Resource Constrained Systems

We propose a distributed control algorithm using a new idea of collaborative processing for networked control systems with limited computational resources. In a networked control system, the processing unit is required to complete multiple tasks (such as control, sensing, signal processing, etc.) in a timely manner. Due to limited computational resources that are available to the processing unit, the assumption that the control input is computed at every time step may no longer hold. In this paper, we consider that there are multiple remote controllers and each controller controls its own plant. We give the controllers within a neighborhood the ability to compute their neighbors’ control inputs. By suitable design of the processor collaboration algorithm, we are able to ensure the stochastic stability of the entire system even if there are random computation interruptions occurring at different time steps. Numerical results verify the effectiveness of the algorithm.