Search

A fuzzy-PID speed controller proposed in this paper is designed for the Doubly Salient Motor (DSM) with axial flux, which has high efficiency of rotor and power density. This paper started from the structure and mathematical model of axial flux DSM, and then analyzed its operating characteristics. Due to nonlinear functions of Flux and torque rotor position and phase current, application of linear control methods cannot meet the requirements of non-linear, variable structure and variable parameters. In order to stabilize the motor and achieve the maximum speed quickly, a fuzzy-PID speed controller utilizes the fuzzy control strategy to get tuning rules of three parameters in PID controller, comparing to very complicate conventional PID control parameters self-tuning process. Simulation results show that the fuzzy PID control is superior to the traditional PID control, which can effectively improve the response speed and interference immunity, shorten the rise time and reduce the overshoot.

Kalman filter is widely applied in engineering due to its real time recursive algorithm, while it is hard to implement hardware with traditional device in some actual systems with hign performances in real time and in flexible and convenient implementation. Applying plentiful resources of FPGA and its flexibility and parallel computation, according to the "first timing later circuit" mind, this paper designed a floating Kalman filter by using the top-down synchronous design method based on FPGA. The design uses time-sharing multiplex technology in IP cores and modules so as to save resources and to implement hardware conveniently on the premise that the system has met the real-time requirement. Taking the precision improvement of the GPS as its application, the Modelsim simulation results show that the designed floating Kalman filter is real time and easy to realize and simultaneously saves hardware resources, so it has practical value.

Kalman filter is widely applied in engineering as its realtime recursive theory. Traditional design flow to implement the filter has problems of repetive description and hidden mistakes, because system is programmed again after the simulation and verification in Matlab environment. The paper proposes a novel approach to realize Kalman filtered system, in which designs system from Matlab model directly to FPGA hardware. That is, model and simulate the filter system in Matlab, and generate hardware description code after the verification without need of hardware program manually. This approach has the advantages in the flexibility of the FPGA, and in the efficiency of the system implement, moreover in the consistency between results and simulation. The approach is suitable for any algorithm and models which can be described in Matlab, and will be applied widely.