Accurate and versatile multivariable arbitrary piecewise model regression of nonlinear fluidic muscle behavior

Accurate multivariable arbitrary piecewise model regression of McKibben and Peano muscle static and damping force behavior

“…MAPMORE is an algorithm that uses experimental data to build multivariable piecewise functions. It was introduced by Veale et al [35] for the accurate modeling of McKibben and Peano muscle quasistatic force behavior. MAPMORE selects the mathematical modeling terms that constitute its piecewise functions from a dictionary of user specified terms and fits the terms to the data.…”

“…This involves extraction, segmentation, and scaling of training data; choice of suitable model terms; and multivariable piecewise function fusion and extrapolation. In this work, the versatility of MAPMORE in accurate model generation is extended beyond the quasistatic force model generation of Veale et al [35] to dynamic situations where the damping force component of fluidic muscles is also modeled. Figure 3 and the explanation that follows outline how MAPMORE works when applied to generating models of a fluidic muscle's damping force component.…”

“…Note that the symbols in the MAPMORE algorithm referred to in this explanation are summarized at the end of this section in table 2. Also note that the process used to generate the static force component models in this paper is similar, and is described in detail in [35]. In the first step in MAPMORE, in (a) of figure 3, data is collected.…”

“…From the vector of measured force data F the training data damping force component F dt is calculated (c). Using the force component predicted by the static force model F s (see [35] for details), an uncorrected training component F ′ dt is calculated as F ′ dt = F − F s . Then, steadystate static force modeling errors are accounted for by subtracting the steadystate value of F ′ dt between the rise and fall segments of each experiment, F dt,ssv , from F ′ dt :…”

“…Specifically this paper presents the validation of Multivariable Arbitrary Piecewise MOdel REgression (MAPMORE) generated static and damping force models as more accurate than existing accurate models for Peano and McKibben muscles. It also builds on our past work's application of MAPMORE to static behavior [35] by presenting the validated application of MAPMORE to generating muscle damping force models.…”