Gait Impairment Score: A Fuzzy Logic-Based Index for Gait Assessment

The objective assessment of subject’s gait impairment is a complicated task. For this reason, several indices have been proposed in literature for achieving this purpose, taking into account different gait parameters. All of them were essentially based on the identification of “normality ranges” for the gait parameters of interest or of a “normal population”. However, it is not trivial to obtain a unique definition of “normal gait”. In this study we proposed the Gait Impairment Score (GIS) that is a novel index to evaluate the subject’s gait impairment level based on fuzzy logic. This index was obtained combining two Fuzzy Inference Systems (FISs), based on gait phases (GP) and knee joint kinematics (JK) parameters, respectively. Eight GP parameters and ten JK parameters were extracted from the basographic and knee kinematic signals, respectively. Those signals were acquired, for each subject’s lower limb, using a set of wearable sensors connected to a commercial system for gait analysis. Each parameter was used as input variable of the corresponding FIS. The output variable of the two FISs represented the impairment level from the GP and JK point of view. GP-FIS and JK-FIS were applied separately to both right and left leg parameters. Then, the fuzzy outputs of the two FISs were aggregated, independently for each side, to obtain the leg fuzzy output. The final subject’s GIS was obtained aggregating the fuzzy outputs of the two legs. The score was validated against two gait analysis experts on a population of 12 subjects both with and without walking pathologies. The Analytic Hierarchy Process (AHP) pairwise comparisons were used to obtain the subjects’ ranking from the two experts. The same population was scored using the GIS and ordered in ascending order. Comparing the three rankings (from our system and from the two human experts) it emerged that our system gives the same “judgment” of a human expert.