A Novel Analog Multisensor Design Based on Fuzzy Logic: A Magnetic Encoder Application

This paper proposes a new multisensor data fusion methodology based on fuzzy logic for a high precision magnetic encoder design. The proposed multisensor design uses a projected electronic circuit to implement analogically the Takagi-Sugeno model, maintaining the measuring signal continuous nature. This method differs from others due to its new approach regarding the typical fuzzification interface, as it uses each sensor characteristic curve as the fuzzy system membership function. The developed interpolation technique evaluates the angular position membership degree to each of the magnetic encoder hall-effect sensors, allowing the fuzzy encoder to fully estimate the intermediary positions between the sensor elements and, thus, present a linear and continuous output. The fuzzy encoder is completely implemented on hardware, dismissing any software interface and, consequently, reducing the encoder complexity and cost, while improving the frequency response. Computer simulations and experimental results on a prototype are presented to validate the fuzzy encoder performance. The proposed methodology and hardware design, although applied on a magnetic encoder, are applicable to others spatial measurement systems based on detectors.