Resumen

CABRERA SARMIENTO, Alejandro José; SANCHEZ-SOLANO, Santiago  y  GARCIA GUIROLA, Yasmani. Fuzzy inference module with variable knowledge base on reconfigurable hardware. EAC [online]. 2021, vol.42, n.2, pp. 34-54.  Epub 09-Sep-2021. ISSN 1815-5928.

This paper presents the development of a fuzzy inference module (FIM) implemented on reconfigurable hardware with the ability to dynamically modify its knowledge base. The original FIM is designed using the fuzzy systems development environment Xfuzzy, which allows the generation of VHDL hardware description language code for the FIM architecture. Subsequently, the VHDL code is modified to add ports with the address, data and read / write control signals to the antecedents and rule memories that contain the knowledge base. The modified FIM is encapsulated in an intellectual property module following two possible options, making the corresponding interconnections in each case, so that these memories can be accessed through the added port from a processing system embedded in the same device, enabling the modification of its contents in operation time. The implementations were carried out on two types of reconfigurable devices: a Spartan-3E1600 FPGA, using a processing system based on the Microblaze softcore and the ISE/EDK development environment; as well as on a SoC-FPGA Zynq-7Z010, using its hardcore processing system based on ARM and the Vivado development environment, checking the dynamic modification of the FIM knowledge base. The carried out modifications make it easy to fine-tune the knowledge base during the development stage of a hybrid hardware/software fuzzy controller as well as the implementation of an adaptive fuzzy controller.

Palabras clave : Fuzzy inference module; Xfuzzy; FPGA; SoC-FPGA.

        · resumen en Español     · texto en Español     · Español ( pdf )