A unified tool for fuzzy/neural network systems

Successful designers of complex control systems use a hierarchical approach, breaking down a problem into parts, or objects. Connecting the objects is another simple problem on the way to a reliable system description with known behaviors and internal structure. Though many systems implement traditional control and use specialized neural network and fuzzy inference software, this article introduces an integrated development environment. Designers can use this unified fuzzy logic-neural network tool set called IDEA (Integrated Development Envelopment for Autonomous Response Systems) to develop a complex neurofuzzy controller. The tool set includes a multiple document interface to edit EDA-language text files, a C++ library of a biology-based model, graphic objects for the design phase, an EDA-language compiler called CARS, and tools to produce either a software simulation or a specific electronic circuit. The high-level description language called EDA lets designers implement the description to develop control systems or support decision systems, interconnecting well defined internal fuzzy structures and neural or traditional algorithms. Designers can use their own interface to a C++ class library, a high-level description language, and a development environment for these elements and several tools.