Can Micro-Doppler Predict Human Movement?

Withdrawn

• Conditions: Pain Management; Injury Prevention

• Registration Number: NCT04468828
• Lead Sponsor: Milton S. Hershey Medical Center

Brief Summary

The analysis of human motion using radar has become an increasingly active topic of study due to the diverse applications offered by such an analysis (Lai et al., 2008; Fairchild \& Narayanan, 2016; Narayanan et al., 2014). Information about human motion has important applications for urban military operations, search-and-rescue missions, surveillance, and hospital patient monitoring. The micro-motions of human movement in the presence of radar illumination creates unique modulations in the received signal known as the micro-Doppler effect. By analyzing these frequency modulations, one can infer the type of movement being performed. This micro-motion associated with human movement produces a nonlinear and non-stationary signal that can be characterized using time-frequency domain analysis. Such signals will be used to identify high injury risk versus low injury risk athletes, which creates an opportunity to direct limited prevention resources to these high-risk athletes; identify individuals at risk of falls; and, may even be useful in diagnosing conditions such as Parkinson's where asymmetrical movement patterns occur as an early indicator.

Traditional methods of movement analysis involve the use of expensive video motion capture systems that accurately measure the 3-dimensional position of passive reflective markers affixed to human body landmarks such as joints and body segments, and while motion capture systems are used to effectively estimate movement dynamics, they are generally not portable, they are expensive, and they can be cumbersome when the reflective markers are applied to older persons or persons with movement deficiencies. Drs. Narayanan and Onks have successfully tested a novel use of Doppler radar that is portable, less expensive, and eliminates the need for affixing cumbersome reflective markers to participants. In addition, preliminary testing has demonstrated the ability to discriminate between certain movement conditions at a level of precision we feel are not obtainable with video motion capture.

Detailed Description

Each subject will report for their scheduled data collection in the biomechanics laboratory. Boney landmarks (shoulder, hip, knee, ankle, etc.) will be used to place small reflective markers for use with motion capture analysis. The radar will be positioned so that the MDS data can be captured simultaneously with the motion capture data. Each volunteer will complete the following activities: • Walk with athletic shoes

* Walk with bilateral heel lifts in shoes

* Walk with unilateral heel lift

* Squat jump with athletic shoes

* Squat jump with bilateral heel lifts in shoes

* Squat Jump with unilateral heel lift

* Stationary standing posture with athletic shoes

* Stationary standing posture with bilateral heel lifts in shoes

* Stationary standing posture with unilateral heel lift Each activity will be performed three times toward the radar in order to assess repeatability and reliability, and also for maintaining adequate statistical training and testing datasets for confirming the use of previous established classification algorithms. These algorithms will again be used to calculate prediction accuracy for the different activities (walking, jumping, and stationary posture) and different footwear conditions (shoes without inserts, shoes with bilateral inserts, shoes with unilateral inserts). The biomechanical motion capture data will be processed similarly to the MDS data to compare to accuracy between the two methods for the same observed task-specific differences. Our goal is to determine if MDS can achieve the same measurement accuracy as motion capture for the same task. Since biomechanical motion capture is the "gold standard" of human movement measurement, the successful completion of this aim will establish the validity of MDS as an effective clinical measure of human movement.

Study Timeline

• Study First Submitted: 2020-07-09
• Study First Submitted QC: 2020-07-09
• Study First Posted: 2020-07-13
• Study Start: 2021-07
• Status Verified: 2021-11
• Last Update Submitted: 2021-11-03
• Last Update Posted: 2021-11-10
• Primary Completion: 2022-01
• Study Completion: 2022-01

Recruitment & Eligibility

• Status: WITHDRAWN
• Sex: All
• Target Recruitment: Not specified

Inclusion Criteria

• Adults aged 18-25
• Able to effectively read, write and understand English
• Must be able to walk, jog, and jump without a limp

Exclusion Criteria

• Children 0-17 years of age
• Unable to effectively read, write and understand English
• Subjects with conditions that limit the ability to walk, jog, or jump
• History of hip or knee surgery
• History of cerebral vascular accident
• Subjects who are pregnant

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
The primary endpoint of this proposed study will compare MDS to biomechanical motion capture in terms of accuracy of discriminating footwear.	40 minutes	We expect that when the motion capture data is processed similarly to the MDS data, the predictive accuracy of the MDS approach will be at least as good as that of the motion capture approach.

Secondary Outcome Measures

Name	Time	Method
To classify different human movement features by pairing different types of human movement to their component features.	40 minutes	This analysis will help to establish a radar MDS "library" of features matched with specific types of movements.