![]()
How Do I Design Using Fuzzy Logic?
- Fide helps you boost your productivity
- Fide helps you accomplish improved system performance
- Fide enables you to provide a lower cost solution
- Fide helps you bring products to market faster
You do not need to be an expert in mathematics or fuzzy set theory in order to develop a fuzzy controller. All you need is a development tool which supports a well defined design methodology. The sequence of steps required to design a fuzzy controller is illustrated in Figure 9:
First, we understand the physical system, its control requirements, and identify the controller's inputs and outputs. Second, we define the ranges and labels for the membership functions and describe the controller's operation using fuzzy rules. Finally, we debug and tune the controller by modifying membership functions, or rules, whenever appropriate.
Aptronix has created Fide, the Fuzzy Inference Development Environment an easy to use integrated tool set which supports all phases of application development, concept, design, tuning, simulation, and implementation.
Fide incorporates some advanced capabilities which will enable you improve productivity and system performance while reducing hardware costs and accelerating time to market:
Fide helps you boost your productivity
Fide has been created with the application engineer in mind. It supports a design philosophy which is similar to traditional methods, but much simpler because it eliminates the tedious mathematical modeling by taking care of it internally. With Fide you do not need be a programming expert since you can describe your system in simple English-like rules. Moreover, Fide automatically generates very compact code for your target platform. As a result, Fide helps you focus more on developing your application rather than programming. Fide also provides an easy to use graphical interface, as well as breakthrough debugging, and tuning capabilities which can help you achieve unprecedented levels of productivity.
Fide helps you accomplish improved system performance
Fide includes some unique algorithms to implement the Fuzzy Inference Process into conventional microcontrollers. For example, the labels for each membership functions require only a couple of bytes to represent. In addition, Aptronix's proprietary algorithm called the Truth Value Flow Inference (or TVFI) built into Fide can speed up fuzzy inference and reduce object code significantly. Customer benchmarks comparing Fide with other tools have shown that TVFI can accelerate the inference process as much as 10 to 100 times and reduce object code as much as 5 to 10 times. The processing speed and code size combined will allow you to add more features in your low end applications or improve speed in the high end which results in overall better product performance.
Fide enables you to provide a lower cost solution
In many applications where cost is an important differentiator, Fide can prove to be an invaluable companion to you. With its compact code generator and faster inference process, Fide helps you enjoy tremendous hardware cost savings in memory, processor power, or both. For example, the 2-stage inverted pendulum occupies only 950 bytes of memory and runs on a low cost 8-bit MCU. Not only does this result in lower hardware costs, but also development costs assuming we can use existing hardware to implement the fuzzy solution.
Fide helps you bring products to market faster
Real life control requires the design of several sub-systems including a model of the plant. Fide provides a unique capability which allows you to break up large designs into smaller segments. Different modules can be graphically interconnected to compose a larger system. Fide's modular approach allows you to manage very complex system quickly and easily which can help you speed up time to market.
Consider our temperature controller example. Suppose we wish to implement a high end model which includes three different modes, normal, energy saver, and fast. Using other development tools three different controllers would have to be designed. Alternatively, Fide's modularity allows you to replicate the original temperature controller and simply map the domain of the membership functions for the additional modes. This helps you deliver a product variation much faster into the marketplace and cope effectively with today's competitive pressures.
© 1996 Aptronix Inc.