Background: The information of electromyographic signals can be used by\r\nMyoelectric Control Systems (MCSs) to actuate prostheses. These devices allow the\r\nperforming of movements that cannot be carried out by persons with amputated\r\nlimbs. The state of the art in the development of MCSs is based on the use of individual\r\nprincipal component analysis (iPCA) as a stage of pre-processing of the classifiers. The\r\niPCA pre-processing implies an optimization stage which has not yet been deeply\r\nexplored.\r\nMethods: The present study considers two factors in the iPCA stage: namely A (the\r\nfitness function), and B (the search algorithm). The A factor comprises two levels,\r\nnamely A1 (the classification error) and A2 (the correlation factor). Otherwise, the B factor\r\nhas four levels, specifically B1 (the Sequential Forward Selection, SFS), B2 (the Sequential\r\nFloating Forward Selection, SFFS), B3 (Artificial Bee Colony , ABC), and B4 (Particle\r\nSwarm Optimization, PSO). This work evaluates the incidence of each one of the eight\r\npossible combinations between A and B factors over the classification error of the MCS.\r\nResults: A two factor ANOVA was performed on the computed classification errors\r\nand determined that: (1) the interactive effects over the classification error are not\r\nsignificative (F0.01,3,72 = 4.0659 > fAB = 0.09), (2) the levels of factor A have significative\r\neffects on the classification error (F0.02,1,72 = 5.0162 < fA = 6.56), and (3) the levels of\r\nfactor B over the classification error are not significative (F0.01,3,72 = 4.0659 > fB = 0.08).\r\nConclusions: Considering the classification performance we found a superiority of\r\nusing the factor A2 in combination with any of the levels of factor B. With respect to the\r\ntime performance the analysis suggests that the PSO algorithm is at least 14 percent\r\nbetter than its best competitor. The latter behavior has been observed for a particular\r\nconfiguration set of parameters in the search algorithms. Future works will investigate\r\nthe effect of these parameters in the classification performance, such as length of the\r\nreduced size vector, number of particles and bees used during optimal search, the\r\ncognitive parameters in the PSO algorithm as well as the limit of cycles to improve a\r\nsolution in the ABC algorithm.
Loading....