Recently, there has been many practical applications of fuzzy control. Applications of fuzzy control to vehicles are more effective. However, they require more input and output channels as well as higher speed inference ability than typical applications require.
We have developed a tool for developing a fuzzy control system available on a vehicle. The tool consists of a fuzzy inference board and a laptop personal computer with a color display executing a user interface software. The fuzzy inference board can be directly connected with an electronic control unit for vehicle control, and has 10 input channels, 5 output channels and 256 rules. The user interface software mainly provides two functions:
-
1)
The screen editor for membership functions and rules.
-
2)
The real-time monitor for input, output, and grade values when fuzzy inference is executed.
The development tool reported in this paper was used in the development of fuzzy steering control in an automatic driving system in a test course.
Many applications have been publicized for fuzzy control in various fields in recent years. Most applications, including those for consumer electronic appliances, have a small number of input/output channels and rules for fuzzy inference, and high speed processing is not an important requirement. For these reasons, majority of these applications use the technique of software processing or look-up table method using a microprocessor [1]*.
Application of fuzzy control to vehicles enables the driver to realize a control close to human sense. Actually, however, control of an vehicle is a real-time control which requires high-speed fuzzy inference and a large number of input/output channels and rules. For these reasons, conventional fuzzy inference techniques are hardly useful for the control of vehicles.
Membership functions and rules are desirable to be tuned according to the result of monitoring the execution of fuzzy inference while controlling a vehicle practically. Conventional development tools for fuzzy control systems have a number of problems if they are to be installed on a vehicle, such as difficulty of connecting the tool to an electronic control unit (ECU) on a vehicle [2][3][4].
We have developed a tool for developing a fuzzy control system capable of executing fuzzy control, monitoring the execution of fuzzy inference, and editing membership functions and rules on a vehicle, thus facilitating experiment of fuzzy control on vehicles [5].
The development tool reported in this paper was used in the development of fuzzy steering control in an automatic driving system in a test course [6].