The zero-time-controllability of an observer-based deadbeat controller (DBC) is described in this paper. Deadbeat control is a typical example of linear control strategies in discrete-time systems. The standard objective of deadbeat control is to drive any initial state of the system to zero in the shortest time possible. If the entire state is available for direct measurement, this control objective can be accomplished by linear state-variable feedback, which is constant and independent of the initial state. However, if some state variables are not accessible, direct-measurement-based control cannot be implemented. A common solution for dealing with the inaccessible state variables is to introduce an observer that can reconstruct the actual state from measurable data in the shortest time possible, and then apply the state-variable control law. The observer characteristics are independent of this initial state. The additional dynamics introduced in the reconstruction of the state inevitably prolong the control process. Under the additional dynamics, it is still possible for observer-based DBC to satisfy the deadbeat requirement for every initial state of the system.

This paper is a continuation of an earlier one [1]. The observer-based DBC means that the values of the target variables are sure to become zero when only a subset of the entire set of the target variables is accessible to the control. The accessibility to the full set and partial set of the target variables is referred to as partial interconnection and full interconnection, respectively. This paper focuses on discussing two issues. First, the partial interconnection is equivalent to the full interconnection if the values of the inaccessible target variables can be reconstructed from the values of the accessible target variables. Second, finding the observer-based admissible DBCs is equivalent to finding the admissible DBCs if the inaccessible target variables can be reconstructed.

Overall, this paper describes the above-mentioned two issues at length. It is less readable than the authors’ 2012 paper [1]. Although they tried to extend their previous work [1], it is not well described. From my perspective, I can barely see any major distinction between the two papers.