Elsevier

Journal of Biomechanics

Volume 47, Issue 7, 7 May 2014, Pages 1577-1581
Journal of Biomechanics

An individual and dynamic Body Segment Inertial Parameter validation method using ground reaction forces

https://doi.org/10.1016/j.jbiomech.2014.03.004Get rights and content

Abstract

Over the last decades a variety of research has been conducted with the goal to improve the Body Segment Inertial Parameters (BSIP) estimations but to our knowledge a real validation has never been completely successful, because no ground truth is available. The aim of this paper is to propose a validation method for a BSIP identification method (IM) and to confirm the results by comparing them with recalculated contact forces using inverse dynamics to those obtained by a force plate. Furthermore, the results are compared with the recently proposed estimation method by Dumas et al. (2007). Additionally, the results are cross validated with a high velocity overarm throwing movement. Throughout conditions higher correlations, smaller metrics and smaller RMSE can be found for the proposed BSIP estimation (IM) which shows its advantage compared to recently proposed methods as of Dumas et al. (2007). The purpose of the paper is to validate an already proposed method and to show that this method can be of significant advantage compared to conventional methods.

Introduction

Calculating Body Segment Inertial Parameters (BSIP) has been shown to be of critical importance for clinical and biomechanical research (Andrews and Mish, 1996, Kingma et al., 1995, Silva and Ambrósio, 2004, Rao et al., 2006, Pai, 2010). The measurement of the inertia and the position of the COM of each body part allows us to monitor variations in muscle-mass during hospitalization, rehabilitation or neurological examination. Consequently the knowledge of the individual inertial parameters is of crucial importance to support personalized healthcare. The better the inertial estimation of those segments is, the better are the resulting joint loads (force and moment) obtained by inverse dynamics (Pearsall and Costigan, 1999, Rao et al., 2006, Pàmies-Vilàa et al., 2012). Previous research has been conducted to improve the BSIP estimation using geometric models based on numerous anthropometric measurements (Hanavan Jr., 1964) or taking results from cadavers׳ studies (Dempster, 1955, Chandler et al., 1975) as well as in vivo body scanning methods (Zatsiorsky and Seluyanov, 1983, Zatsiorsky et al., 1990, de Leva, 1996, Ma et al., 2011). Even though the estimations have been improved (Dumas et al., 2007) there are still regression methods based on earlier collected databases e.g. (McConville et al., 1980, Young et al., 1983). Geometrical methods are precise and based on complex acquisition systems such as 3D scanner, IRM or X-ray absorptiometry which are expensive and may expose subjects to radiations. Recently, identification methods used in Mechanical Engineering have been applied to the estimation of human BSIP (Atchonouglo et al., 2008, Venture et al., 2009b, Venture et al., 2009c, Ayusawa et al., 2011). These methods are based on human body mechanical models whose parameters are expected to match kinematic and dynamic recorded data. Therefore, they allow evaluating BSIP on a subject-by-subject basis using an optoelectronic motion capture system and a force platform.

In this framework, this paper proposes to evaluate an identification method (IM) to assess the inertial parameters of humans without considering joint torques. It is based on the fact that the dynamics of such systems can be written using the Newton–Euler formalism for the base-link (chosen arbitrarily) and the Lagrangian formalism for the rest of the kinematic chains (Venture et al., 2009b, Venture et al., 2009c, Ayusawa et al., 2011). It can be thus, demonstrated that to identify the dynamics of the whole system only the six equations obtained for the base-link are necessary. Since no ground truth value of the BSIP is available, the IM is validated twofold. First the recalculated contact forces, using inverse dynamics, are compared with the ground reaction forces (GRF) measured from a force platform. Secondly the results are compared with the GRF computed using a regression based model (RM) proposed by (Dumas et al., 2007). Additionally, the results are cross validated with a high velocity overarm throwing movement.

The paper is structured as follows: in the Methods section the obtained identification model from the base link equation is briefly described and the experimental identification of the whole-body parameters is presented. In the Results section, the experimental results obtained from both methods are presented and discussed.

Section snippets

Methods

To obtain accurate identification results it is important to define the kinematic model used to describe the human body, and to obtain its characteristic geometric parameters.

Results

In this section we present the results of the comparison between IM and RM. The RMSE, the Pearson Correlation Coefficient, the phase dependent (P) and magnitude dependent (M) and combined error (C) computations are shown in Table 2. The correlation between the IM and the force plate is very high (>0.99) and the RMSE and P, M and C are smaller compared to the obtained results by RM (Table 3).

Further investigations of each force plate output show that the highest correlations for both BSIP

Discussion

The present study was aimed to validate and compare an identification method of subject specific BSIP estimation with results from a regression based model (Dumas et al., 2007). The BSIP estimation strongly depends on the segment׳s length, mass and geometric shape. During gait, errors in BSIP values have significant effects on kinetic parameters e.g. variations of the hip flexion/extension torque up to 20% depending on the chosen model (Rao et al., 2006). These differences seem to be more

Conclusion

This research was conducted to validate a non-invasive and not radiation based method to obtain individual BSIP for each subject. In comparison with the measured ground reaction forces the IM approach has shown its advantages compared to models proposed in previous research (Dumas et al., 2007). Our work does not contradict previous research but simply provides a new approach to compare BSIPs. The technique is not time consuming and DOF and segments can easily be added or removed upon needs.

Conflict of interest statement

No conflict of interest.

Acknowledgments

This study was supported by the Japan Society for the Promotion of Science (JSPS/FF1/391 ID no. PE 12509). Special thanks to Prof. Yoshihiko Nakamura and Dr. Ko Ayusawa from the University of Tokyo for technical discussions and to Yusuke Ogawa for his kind help during the data acquisition.

References (40)

Cited by (25)

  • An extended statically equivalent serial chain—Identification of whole body center of mass with dynamic motion

    2021, Gait and Posture
    Citation Excerpt :

    We found that it was possible to estimate a set of subject-specific SESC parameters capable of reproducing both ground reaction forces and CoP displacement for a series of postures. When compared to a full dynamic identification such as the ones presented in [12–14,32] the SESC method may offer one advantage: ease of implementation. The full dynamic identification model, as stated in the introduction, requires large accelerations of the body segments in order to obtain a suitable excitation of the relevant parameters.

  • Predictive regression modeling of body segment parameters using individual-based anthropometric measurements

    2019, Journal of Biomechanics
    Citation Excerpt :

    For brevity purposes, only a summary is provided here. These BSP estimation approaches include regression equations from cadaver data (Chandler et al., 1975; Dempster, 1955), imaging techniques (de Leva, 1996), geometric modeling of the body (Pavol et al., 2002), inverse dynamics analyses (Hansen et al., 2014), static force plate analyses (Chen et al., 2011; Damavandi et al., 2009) and photographic analysis (Jensen, 1978; Sanders et al., 2015). Methods have also been developed utilizing individual anthropometric measurements in order to predict whole body COM position (Erdmann and Kowalczyk, 2015), as well as BSPs of all major body parts (Hatze, 1980).

  • Age and body mass index associations with body segment parameters

    2019, Journal of Biomechanics
    Citation Excerpt :

    The results of this study likely provide more representative segment parameters in individuals due to actually measuring the masses of each of the slices. While methods using optimization algorithms to determine parameters from inverse dynamics (Hansen et al., 2014) and static positioning on force plates (Chen et al., 2011; Damavandi et al., 2009) can estimate segment parameters in a non-invasive manner, they are somewhat limited to predefined anthropometric sets due to the placements of visual markers, and assumptions regarding these marker locations relative to anatomical axes and rotation centers. DXA scanning methods have the advantage of determining the masses of each pixel in the image, and determining segment boundaries based on specific anatomical landmarks, allowing more precise segment boundaries, which may be altered depending on the desired anthropometric data set and application (Merrill et al., 2018).

View all citing articles on Scopus
View full text