Use of Wearable Sensors to Assess Prosthetic Alignment in Veterans with Unilateral Transtibial Amputations

Veterans with transtibial amputation require a prosthesis to walk and are at an increased risk of secondary

injury, discomfort, and reduced quality of life. Proper prosthetic alignment can reduce these risks and improve

functional ability and comfort in individuals with transtibial amputation. Currently, prosthetists (clinicians)

subjectively align a prosthesis, and this may require several clinical visits. Subjective alignment relies on

prosthetists’ experience and visual inspection of walking, which is prone to errors and time consuming. Thus,

there is an urgent need to develop objective tools for prosthesis alignment. We aim to develop a novel method

to [assess] prosthesis alignment accurately, precisely, and cost-efficiently using wireless sensor technology,

which could improve quality of life and reduce secondary injury risk for the millions of prosthesis users in the

United States. The goal of this study is to determine the accuracy and precision of using wearable sensors

combined with an algorithm to assess prosthesis alignment in 10 Veterans with transtibial amputation.

[We will ask 10 Veterans with transtibial amputation to walk on a force-treadmill at 1.25 m/s while they use

a prosthesis with neutral alignment and that varies by 3° and 6° in one of three planes, the sagittal, coronal,

and transverse planes, for a total of 13 prosthesis alignments. For each alignment condition, we will determine

the accuracy and precision of using inertial measurement units (IMUs) combined with a numerical algorithm

to estimate dynamic-to-static angle (DSA) of the prosthesis and the biological shank during walking in 10

Veterans with unilateral transtibial amputation (Aim 1), where DSA provides information regarding the

orientation of the prosthesis and the biological shank. We will also determine the accuracy and precision of

using IMUs combined with a numerical algorithm to estimate inter-limb symmetry indices of step length, step

frequency, and contact time, which are important discrete temporal-spatial parameters during walking in 10

Veterans with unilateral transtibial amputation (Aim 2). We will compare results estimated using IMUs with

results calculated using traditional gold-standard measurements of 3D motion capture and ground reaction

forces. We will also investigate the association between angular changes in prosthesis alignment and DSA

and interlimb symmetry indices (Aim 3).] We hypothesize that the IMU method will provide accurate (root-

mean squared error [RMSE]0.75) estimations of

DSA for both legs and inter-limb symmetry indices (mean absolute percentage error [MAPE]0.75)

of temporal-spatial parameters. We also hypothesize that changes in prosthesis alignment will result in

significant differences in DSA using the IMU method and motion capture measurements. We hypothesize

that changes in prosthesis alignment will result in significant differences in interlimb symmetry index of step

length, step frequency, and ground contact time. [If our results suggest that the IMU method does not provide

accurate and precise estimations of DSA or symmetry indices, we will consider using a more sophisticated

prediction model (e.g. machine learning) to predict DSA using IMUs.]

The outcome of our research is the development of a novel method that uses wireless IMU sensors to

[assess] prosthesis alignment accurately (comparable accuracy with camera-based motion capture system

and force plate ground reaction force system), quickly (within a single visit), and cost-efficiently (