Impact of model geometry and joint center locations on inverse kinematic/dynamic predictions: A comparative study of sexually dimorphic models.

This work illustrates the sensitivity of demographically characteristic body segment inertial properties and subject-specific customization on model performance. One characteristic demographic, gender, and one subject-specific characteristic, hip joint center location, were represented with musculoskeletal modeling to evaluate how design decisions may alter model outputs. Generic sexually dimorphic musculoskeletal models were developed from the commonly used Rajagopal model using male and female data adapted by Dumas et al. Hip joint centers of these models were adjusted based on functional joint center testing. The kinematics and dynamics of 40 gait cycles from four subjects are predicted using these models. Two-way analysis of variance (ANOVA) was performed on the continuous time series data using statistical parametric mapping (SPM) to assess changes in kinematics/dynamics due to either choice in model (Rajagopal vs Dumas) or whether joint center adjustment was performed. The SPM based two-way ANOVA of the inverse dynamics found that differences in the Rajagopal and Dumas models resulted in significant differences in sagittal plane moments during swing (0.115 ± 0.032 Nm/kg difference in mean hip flexion moment during initial swing and a 0.077 ± 0.041 Nm/kg difference in mean hip extension moment during terminal swing), and differences between the models with and without hip joint center adjustment resulted in significant differences in hip flexion and abduction moments during stance (0.217 ± 0.055 Nm/kg increased mean hip abductive moment). By comparing the outputs of these differently constructed models with each other, the study finds that dynamic predictions of stance are sensitive to positioning of joint centers, and dynamic predictions of swing are more sensitive to segment mass/inertial properties.

DeepBBWAE-Net: A CNN-RNN Based Deep SuperLearner for Estimating Lower Extremity Sagittal Plane Joint Kinematics Using Shoe-Mounted IMU Sensors in Daily Living.

Measurement of human body movement is an essential step in biomechanical analysis. The current standard for human motion capture systems uses infrared cameras to track reflective markers placed on a subject. While these systems can accurately track joint kinematics, the analyses are spatially limited to the lab environment. Though Inertial Measurement Units (IMUs) can eliminate these spatial limitations, those systems are impractical for use in daily living due to the need for many sensors, typically one per body segment. Due to the need for practical and accurate estimation of joint kinematics, this study implements a reduced number of IMU sensors and employs a machine learning algorithm to map sensor data to joint angles. Our developed algorithm estimates hip, knee, and ankle angles in the sagittal plane using two shoe-mounted IMU sensors in different practical walking conditions: treadmill, overground, stair, and slope conditions. Specifically, we propose five deep learning networks that use combinations of Convolutional Neural Networks (CNN) and Gated Recurrent Unit (GRU) based Recurrent Neural Networks (RNN) as base learners for our framework. Using those five baseline models, we propose a novel framework, DeepBBWAE-Net, that implements ensemble techniques such as bagging, boosting, and weighted averaging to improve kinematic predictions. DeepBBWAE-Net predicts joint kinematics for the three joint angles for each of the walking conditions with a Root Mean Square Error (RMSE) 6.93-29.0% lower than the base models individually. This is the first study that uses a reduced number of IMU sensors to estimate kinematics in multiple walking environments.