Summary Research of Thesis

Summary of research is an hardware embedding of the fuzzy logic controller into the FPGA. In the present study examined a pipeline algorithms in order to obtain a compact design and small resource usage.


BAMBANG SISWOYO, NIM: 0522000001, Postgraduate Program University of Brawijaya: Electronic Department, 31 July 2007. Fuzzy Logic Algorithm Implementation in FPGA as Controller of Liquid Mixing Process Temperature Control System. Supervisor: Moch. Rameli, Co-supervisor: Purwanto.

Fuzzy Logic Controller is very advantageous for applications on control systems with unpredictable plants. Many problems related to uncertainties incontrol systems can be solved using Fuzzy Logic Controller. Owing to the progress in FPGA (Field Programmable Gate Array) technology, It is now possible to integrate and embed a Fuzzy Logic Controller into a single chip. A control algorithm applied on a fuzzy logic controller using FPGA technology will result in a control system with high speed processing, because the data processing is accomplished on the hardware-level.

In this research, the design and implementation of a fuzzy logic control algorithm using FPGA has been carried out to control temperature in Liquid Mixing Process. In this implementation, FPGA XILINX Spartan 3 XC3S1000 has been used. The Mamdani Fuzzy Logic method has been applied using singletone output membership function. There has been used two input membership functions, i.e E (Error) and CE (Change Error), both as triangular membership functions. The maximum number of fuzzysets that can be processed is sixteen. The overlapping function is not limited because there have been 256 if-then rule available in a form of  table.

The implementation of fuzzy logic algorithm using FPGA is satisfying as Controller of Liquid Mixing Process Temperature Control System. The results showed that the controller achieved a process speed of 65,4µS, which is equivalent to a maximum sampling frequency of 15.290KHz.

The use of block system in FPGA gave the following data: Slice FlipFlops needed are 3843 or 25% of 1530 availability,  4 input LUT needed are 2398 or 15% of 15360 availability, Ram Blocks needed are 3 or  12% of 24 availability,  MULT18x18s needed are 2 or 8% of 24 availability, GCLKs needed are 2 or 25% of 8 availability, Bonded IOBs needed are 20 or 11% of 173 availability. For further development, fuzzyset variations like S-function, π-function, and others can be added into the hardwares. In order to make easier the membership function modification without re-compilation, the membership function is stored in the external storage like a flashdisk. In addition, the output membership function should be changed into non-singletone.

Cincopa WordPress plugin

YouTube Preview Image

Last updated by at .

Leave a Reply

Your email address will not be published. Required fields are marked *



You may use these HTML tags and attributes: <a href="" title="" rel=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>