Online Position Control Performance Improving Applying Incremental Fuzzy Logic Controller

Malaysia. Abstract. This paper presents the online control system application for improving the DC motor performance. DC motor widely used in industries and many appliances. For this aim fuzzy logic controller is applied. The type of fuzzy controller use is an incremental fuzzy logic controller (IFLC). The IFLC is developed by using MATLAB Simulink Software and implemented in online position control system applying RAPCON board as a platform. The experimental results produced the best gains of the IFLC are 1.785, 0.0056955 and 0.01 for error gain (GE), gain of change error (GCE) and gain of output (GCU) respectively. Its produce smaller rise time, peak time, 0% overshoot and smaller settling time. Beside that the IFLC response also able to follow the set point. The controller response parameters values are also acceptable. It means that the IFLC suitable to be use for improving the position control system


Introduction
DC motors widely used to improve performance, such as industry, home appliances, robot manipulators etc. because it has high reliability, flexibility and low cost. Most DC motor driver applications are in position control or in speed control systems [1].
Several controllers have been applied for position control system with DC motor as a driver such as Proportional Integral Derivative (PID) [1], Fuzzy Logic [2,3], Neural Network (NN) [4] and etc.
PID controller is commonly applied for controlling motor because of it has simple structure and comprehensive control algorithms [1,2]. Besides, it has been implemented in position control systems, but still suffer from poor performance due to non-linear parameters. The PID controller does not provide satisfactory results when the control parameters, loading conditions and motor itself change [5].
Neural network (NN) controller is another kind of controllers that also has been applied for motor control. In the NN, there is abundant of architecture can be used to perform a variety types of functions. There are kind of neural network with high efficiency and strong function generalizing in terms of learning speed and simplicity of the structure. The NN need increasing the number input NN and add some delay to the input NN to improve the performance [4].
In this paper, the incremental fuzzy logic controller (IFLC) as an intelligent controller is applied. The IFLC does not need knowledge model of system, complex structure, learning process and etc. This controller work is based on the principle of human expert decision making in problem solving mechanism [2,3]. The Mamdani inference as a computational fuzzy inference type is used because its decision is more accurate.

DC Motor Modelling Approach
In armature control of separately excited DC motors, the voltage applied to the armature of the motor is adjusted without changing the voltage applied to the field. Figure 1 describes a separately excited DC motor equivalent model [1,6].

Dynamic Model
Voltage equation of the armature circuit under transient is given by [6] From the dynamics of motor load system When field current is kept constant, flux remains constant. Replacing Keф by a constant Ke, yields For no load previous equation become:

Transfer Function
Based on Laplace transforms of Eq. (1) and Eq. (6), the armatur current of motor is Index n refers to the time instant. By tuning we shall mean the activity of adjusting the parameters , , and in order to achieve a good closed-loop performance.
The IFLC as in Figure 2 is of almost the same configuration as the FPD controller except for the added integrator. The conclusion in the rule base is now called change in output (CU), and the gain on the output is accordingly GCU. The control signal ( ) at time instant n is the sum of all previous increments,  The mapping is usually nonlinear, but with the usual favorable choice of design is a linear approximation as Eq. (14). By assuming that * = , = 1 and Td = 0, it is clear that the linear controller is a crisp PI controller. Note that the proportional gain now depends on GCE. The gain 1 ⁄ was determined by the ratio between the two fuzzy input gains, and the inverse of the derivative gain Td in FPD control; the gains GE and GCE change roles in FPD and IFLC controllers.

Online Position Control Implementation
The online position control system with a DC motor as an actuator is implemented using computer, RAPCON board [7] and DC motor.

Experimental Result and Analysis
The experiment for testing the controller performance is done in two ways. There are simulation and online scheme.

Simulation Scheme
The best settings of controller parameters values for simulation scheme are 2.721, 0.019 and0.259 for GE, GCE and GCU respectively. Figure 4 (a) represents the output response of IFLC controller and (b) its zooming.

Online Scheme
The best controller parameter values for online are 1.785, 0.0056955 and 0.01 for GE, GCE and GCU setting respectively.Reference inputs in this case combination between Signal Generator and Constant value. Figure 5 (a) represents the output response of IFLC for online scheme experiment and (b) its zooming. The IFLC responses specifications comparison between simulation and online are elaborated in Table 1.