Challenges and advances in the use of wearable sensors for lower extremity biomechanics

Abstract

The use of wearable sensors for the collection of lower extremity biomechanical data is increasing in popularity, in part due to the ease of collecting data and the ability to capture movement outside of traditional biomechanics laboratories. Consequently, an increasing number of researchers are facing the challenges that come with utilizing the data captured by wearable sensors. These challenges include identifying/calculating meaningful measures from unfamiliar data types (measures of acceleration and angular velocity instead of positions and joint angles), defining sensor-to-segment alignments for calculating traditional biomechanics metrics, using reduced sensor sets and machine learning to predict unmeasured signals, making decisions about when and how to make algorithms freely available, and developing or replicating methods to perform basic processing tasks such as recognizing activities of interest or identifying gait events. In this perspective article, we present our own approaches to common challenges in lower extremity biomechanics research using wearable sensors and share our perspectives on approaching several of these challenges. We present these perspectives with examples that come mostly from gait research, but many of the concepts also apply to other contexts where researchers may use wearable sensors. Our goal is to introduce common challenges to new users of wearable sensors, and to promote dialogue amongst experienced users towards best practices.

Keywords: Gait; Inertial measurement units; Open-source; Real-world; Sensor-to-segment.

Figure 1.

Sagittal plane shank angular velocity from OMC (blue), raw IMC data (solid translucent black), physics-based IMC data (dotted black), and PCA-resolved IMC data (dashed black). The movements were primarily about the flexion-extension axis of the knee.

Figure 2.

Ankle moment (plantarflexion < 0) estimated by tracking simulated gyroscope and accelerometer signals (noise- and bias-free) from only the right shank (blue) during a single gait cycle outperformed estimates from a purely predictive simulation (red) compared to reference data (tracking coordinates and ground reaction forces). Data was taken from the publicly available OpenSim Moco 2D walking example.

Figure 3.

Sensor placement and acceleration trajectories (lumbar sensor) in the horizontal plane for a person with Parkinson’s disease OFF and ON antiparkinsonian medication during quiet standing. (AP - anteroposterior; ML - mediolateral)

Figure 4.

Gait detection and functional reference frame outcome example. A: Detection of probable walking bouts from unobserved data collection. B: 30-stride mean and SD knee flexion excursion calculated about a functional axis from a young adult (blue) and older adult with knee osteoarthritis (red). Note that IMU data in a functional reference frame reveals smaller joint range of motion during stance and throughout the gait cycle for the individual with knee OA.