Fuzzy controller with second order defuzzification algorithm

Fuzzy controllers can handle multiple inputs and outputs however size of the fuzzy table grows exponentially with each input added. In proposed approach two six bits input, (8 /spl times/ 8), (6 /spl times/ 6) and (4 /spl times/ 4) fuzzy look up tables (LUT) were used each having 64, 36 and 16 values respectively. The system uses two inputs, one for the row and the other for the column, the upper three bits of each input are used for addressing the exact position of the nearest data point in the look up table, and the rest of the input is the information about the membership function. Eight adjacent points are used to approximate the required data points to generate the control surface. Second order method proposed in this paper was used to generate the control surface. The algorithm used for defuzzification will be discussed in detail. The advantage of this method over other defuzzification methods like the LUT and the triangular membership function approach is the efficient usage of a relatively smaller look up table, which minimizes the hardware requirement and ensures easier implementation. The error produced using this method is reduced considerably as compared to others; moreover it can also be extend into multiple dimensions easily with minimum number of inputs resulting in smoother control surfaces.