Fitsum E. Petros

Ms. Fitsum E. Petros

Columbia University, Mechanical Engineering

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 May 2021 in Sensors
Verified
Journal Article
Open access
PDF file
MGMichelangelo Guaitolini
Michelangelo Guaitolini
Fitsum E. PetrosFitsum E. Petros
Fitsum E. Petros
APAntonio Prado
Antonio Prado
Angelo Maria SabatiniAngelo Maria Sabatini
Angelo Maria Sabatini
SASunil K. Agrawal
Sunil K. Agrawal
Ageing, disease, and injuries result in movement defects that affect daily life. Gait analysis is a vital tool for understanding and evaluating these movement dysfunctions. In recent years, the use of virtual reality (VR) to observe motion and offer augmented clinical care has increased. Although VR-based methodologies have shown benefits in improving gait functions, their validity against more traditional methods (e.g., cameras or instrumented walkways) is yet to be established. In this work, we propose a procedure aimed at testing the accuracy and viability of a VIVE Virtual Reality system for gait analysis. Seven young healthy subjects were asked to walk along an instrumented walkway while wearing VR trackers. Heel strike (HS) and toe off (TO) events were assessed using the VIVE system and the instrumented walkway, along with stride length (SL), stride time (ST), stride width (SW), stride velocity (SV), and stance/swing percentage (STC, SWC%). Results from the VR were compared with the instrumented walkway in terms of detection offset for time events and root mean square error (RMSE) for gait features. An absolute offset between VR- and walkway-based data of (15.3 ± 12.8) ms for HS, (17.6 ± 14.8) ms for TOs and an RMSE of 2.6 cm for SW, 2.0 cm for SL, 17.4 ms for ST, 2.2 m/s for SV, and 2.1% for stance and swing percentage were obtained. Our findings show VR-based systems can accurately monitor gait while also offering new perspectives for VR augmented analysis.

10 June 2022 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
Fitsum E. PetrosFitsum E. Petros
Fitsum E. Petros
MKMatthew E. Klenk
Matthew E. Klenk
SASunil K. Agrawal
Sunil K. Agrawal
Functional movements require the successful integration of visual, vestibular, and proprioceptive sensory signals. It has been shown that athletes are less affected by external sensory perturbations than untrained individuals. This study aims to investigate if trained athletes respond differently to sensory manipulations during walking compared to non-athletes. A group of Division I football players and non-athletes, 10 subjects each, underwent sensory perturbations while walking overground for 10 laps on a walkway. The sensory perturbations were head yaw oscillations, visual field oscillations, and synchronized head-vision oscillations. A robotic neck brace and a virtual reality system imposed head movements and visual perturbations, respectively. Stride length, stride width, stride velocity, and their respective variabilities were evaluated using a virtual reality system. Results show a condition-dependent change in gait parameters between football players and non-athletes. Visual oscillations caused a wider stride width, shorter stride length, and lower stride velocity in football players when compared to non-athletes.

6 December 2023 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
Fitsum E. PetrosFitsum E. Petros
Fitsum E. Petros
DHDiane Hu
Diane Hu
PKPriya Kulkarni
Priya Kulkarni
SASunil K. Agrawal
Sunil K. Agrawal
Head turns during walking have been used to assess balance, mobility, and vestibular function in both experimental and clinical applications. However, head turns in walking experiments have been limited to self-initiated head motions as opposed to controlled motions. The aim of this study is to evaluate the effects of controlled head turns enabled by a robotic neck brace in elderly and young adults during overground walking under normal (HO) and altered vision (HVO). The robotic neck brace applied controlled sinusoidal head turns around the vertical axis at $\pm 30^{\circ }$ , 0.4 Hz. The vision was altered using a virtual reality headset, where the visual field was aligned along the direction of the head oscillation. Ten elderly (EA, 65–85 yrs) and ten younger (YA, 22–32 yrs) adults were recruited. Spatiotemporal gait parameters, such as stride length (SL), stride width (SW), stride velocity (SV), stride time (ST), stance time percentage (STP), and direction of progression (DoP), as well as mediolateral and anterior-posterior margins of stability ( $ML_{MoS}$ and $AP_{MoS}$ ) were analyzed. Elderly participants showed greater gait changes than younger individuals, particularly during HVO, leading to the highest DoP deviations. Our results indicate that the elderly had difficulty relying on non-visual cues to compensate for the altered vision. However, they had comparable $ML_{MoS}$ and more stable $AP_{MoS}$ . Overall, older adults prioritized balance and stability, while young adults focused on preserving walking direction. Our results indicated that age significantly impacts gait during head oscillations, especially when visual information is altered.