Design of simulation and experimental system implementing fuzzy logic in a programmable device
Stockley, Darleen J.
MetadataShow full item record
Traditional controllers for servo motor systems have been based on motor position-control systems and have been modeled mathematically by digital control systems. For example, the traditional approach for solving dynamic system problems has been to: (1 ) define a system and its components, (2) formulate a mathematical model and list the necessary assumptions, (3) write differential equations describing the model, (4) solve the equations for desired output variables, (5) examine the solutions and the assumptions, and (6) reanalyze or redesign to adjust the desired output variable to preferred values. Utilization of feedback control provided such systems with the ability to adjust the transient and steady-state performance. Clearly, this approach is a highly mathematical approach, and becomes quite complex for systems having higher-order transfer functions. This thesis describes the application of fuzzy logic to the control procedure to provide a less complex control system with a faster adjustment time than that of the traditional control system. Magnitudes of input values and rates of change of input values are input variables for a control system having a second order transfer function. A basic fuzzy logic simulation was developed that included inputs into decision blocks that determined a degree of membership in five membership categories for each of the two input variables, compared and determined the minimum degree of category membership for the two variables, compared and determined a maximum of the values determined for each category and a sum of the maxima, multiplied by the unadjusted category output value, summed the contribution from each category, divided the contribution sum by the maxima sum to provide a feedback value, and subtracted the feedback value from the input signal that alternates between a high and a low frequency. Simulations were run for different pairs of input frequencies, and outputs were examined. An experimental closed loop system was constructed utilizing a binary counter and oscillator, a phase-locked loop, an evaluation board with a microcomputer unit, a digital-to-analog converter, and an operational amplifier. The effectiveness of this system is evaluated.