This paper proposes analysis and synthesis methods of dynamic quantizers for linear feedback single input single output (SISO) systems with discrete-valued input in terms of invariant set analysis. First, this paper derives the quantizer analysis and synthesis conditions that clarify an optimal quantizer within the ellipsoidal invariant set analysis framework. In the case of minimum phase feedback systems, next, this paper presents that the structure of the proposed quantizer is also optimal in the sense that the quantizer gives an optimal output approximation property. Finally, this paper points out that the proposed design method can design a stable quantizer for non-minimum phase feedback systems through a numerical example.
Recently, the fractional order PID (FO-PID) control, which is the extension of the PID control, has been focused on. Even though the FO-PID requires the high-order filter, it is difficult to realize the high-order filter due to the memory limitation of digital computer. For implementation of FO-PID, approximation of the fractional integrator and differentiator are required. Short memory principle (SMP) is one of the effective approximation methods. However, there is a disadvantage that the approximated filter with SMP cannot eliminate the steady-state error. For this problem, we introduce the distributed implementation of the integrator and the dynamic quantizer to make the efficient use of permissible memory. The objective of this study is to clarify how to implement the accurate FO-PID with limited memories. In this paper, we propose the implementation method of FO-PID with memory constraint using dynamic quantizer. And the trade off between approximation of fractional elements and quantized data size are examined so as to close to the ideal FO-PID responses. The effectiveness of proposed method is evaluated by numerical example and experiment in the temperature control of heat plate.
In this paper, we propose a design methodology for on-board diagnosis engine of embedded systems. A boolean function for diagnosis circuit can be mechanically designed from the system dynamics given by the linear differential equation if it is observable, and also if the relation is given between the set of abnormal physical parameters and the faulty part. The size of diagnosis circuit is not so large that it can be implemented in FPGA or fabricated in a simple chip.