Fuzzy logic denoising of the miniaturized inertial sensors in redundant configurations

The paper presents a fuzzy logic based algorithm to reduce the noise of the miniaturized inertial sensors used in strap-down inertial navigation systems (SDINS) in linear redundant configurations. The algorithm idea is to establish a weight for each sensor from an SDINS axis (measuring acceleration or angular speed) as a function of the standard deviation of the last m samples acquired from it by using the fuzzy logic techniques. The chosen fuzzy controller is of type "P" that means it realizes proportionality between the input (standard deviation of the last m samples acquired from sensor) and the output (sensor weight). Three membership functions are considered both for input and for output. Also, the MAX-MIN inference method is used for fuzzyfication, while the centroid method is used for defuzzyfication. The algorithm is implemented in Matlab-Simulink and numerical simulated in two configurations: with four sensors, respectively with nine sensors, in the redundant array mounted on the SDINS axis.