No abstract

Input signals are often fed as analog quantities into fuzzy logic processing systems. This paper proposes a method for configuring interface necessitated by fuzzy logic systems featuring analog inputs and digital processing. This interface is a data converter which processes input analog voltage signals and outputs digital signals suitable for fuzzy logic processing. The conversion principle is based on the concept of level sets within the fuzzyset theory, i.e., depending on whether a given analog signal belongs to level sets, the signal is quantized and converted to membership grade, simultaneously. Having programmable conversion characteristics, it can realize typical membership functions such as Z‐, ‐, and S‐functions. The operation of the converter, which is configured of counters, a memory, a digital‐to‐analog converter, and a comparator, is verified experimentally as regards a non‐monotonic and asymmetric membership function. the conversion accuracy is then discussed and the causes for error are described.

The angular dependence of the coercivity Hc for vacuum evaporated Co–O films with oblique anisotropy is studied both in an in-plane and in a normal plane in order to clarify the magnetization reversal mechanism. The angular dependence of Hc in the in-plane is remarkably different from that in the normal plane. The Co–O films have columnar structure, and the columnar grains look almost the same in both planes. The anisotropy field Hk in the normal plane is much higher than in the in-plane. Simulated curves of the angular dependence of Hc using the curling incoherent magnetization reversal model on the basis of measured values for Hc of the easy axis direction and Hk are similar to the experimental results in both planes. These results indicate that the magnetization reversal mechanism of the Co–O films is the incoherent rotation and that the interaction among the grains is different in both planes.