| dc.description.abstract | Many clinical assessment protocols rely on the evaluation of functional movement tests
such as the Single Leg Squat (SLS), which are often assessed visually. Visual assessment
is subjective and depends on the experience of the clinician. Developing a reliable automatic
human motion tracking and assessment system can improve the accuracy of SLS
clinical assessments and provide objective results that can be tracked and monitored over
time to guide rehabilitation and determine an individual's response to an intervention.
In this study, an Inertial Measurement Unit (IMU) based method for automated assessment
of squat quality is proposed to provide clinicians with a quantitative measure of SLS
performance.
First, an automated pose estimation method is applied to SLS motion data. A set
of three IMUs is used to estimate the joint angles, velocities and accelerations of the
squatting leg. To tackle noisy sensor measurements and gyro drift, a 7 degree of freedom
(DOF) kinematic model of the lower leg was applied together with a constant acceleration
assumption to approximate the angular velocity and linear acceleration at each sensor
location. The kinematic model predictions of the angular velocity and linear acceleration
and sensor measurements were fused via an Extended Kalman Filter (EKF). The position,
velocity, and acceleration of each DOF were defined as the states to be estimated by
the EKF. The pose estimation results showed successful extraction of joint angles with
an average RMS error of 3.2 degrees, 5.5 degrees, 7 degrees compared to joint angles estimated from motion
capture for the ankle, knee, and hip joints, respectively. For this estimation, the required
parameters for the kinematic model, including information about the sensor placement and
orientation as well as the kinematic link lengths, were extracted from the marker data.
However, in clinical applications of the proposed method, when marker data is not
available, these parameters need to be measured. Measuring these parameters is time consuming
in the clinical setting, which limits application of IMUs for clinical purposes. With
the motivation to make this procedure easier and faster, a method for approximating the
parameters using placement assumptions and body measures was described. A sensitivity
analysis was performed to detect those parameters which most affect pose estimation accuracy.
The sensitivity analysis results revealed that sensor orientation is the most critical
factor for accurate pose estimation. In this thesis, a simple and easy to use method is
proposed for sensor orientation calibration, based on a systematic placement of sensors
and using gyroscope information for orientation estimation. This protocol was evaluated
experimentally and pose estimation error with approximated parameters before and after
applying the calibration protocol were compared. The comparison results showed that the
estimate of the sensor orientation increases the pose estimation accuracy by 6.5 degrees for the
knee joint angle and with an average of 1.8 degrees for other joints without the need for time
consuming calibration.
In the second part of the thesis, an algorithm for automated assessment of the SLS
in terms of dynamic knee valgus and risk of knee injury is developed. After applying the
pose estimation algorithm to IMU data of SLS motions, the estimated time series data
of joint angles, velocities and accelerations for consecutive squats were segmented into
individual squat repetitions. Statistical time domain features were generated from each
repetition. The most informative features were selected using a combination of 18 feature
selection techniques. Six common classifiers in including SVM, Linear Multinomial Logistic
Regression, Decision Tree, Naive Bayes, K Nearest Neighborhood, and Random Forests
were applied to the full dimensional data, the subset of selected features, and extracted
features by supervised principal component analysis. The proposed approach was evaluated
in two trials. First, a pilot study was conducted on a small dataset, followed by analysis
on a larger clinical data set, collected by our clinical collaborator. For the clinical study, a
dataset of SLS performed by healthy participants was collected and labelled by three expert
clinical raters using two different labeling criteria: "observed amount of knee valgus" and
"overall risk of injury". Labels included "good", "moderate", and "poor" squat quality
or "high risk", "mild to moderate risk", and "no risk" of injury. Feature selection results
showed that both flexion at the hip and knee, as well as hip and ankle internal rotation
are discriminative features, and that participants with "poor" squats bend the hip and
knee less than those with better squat performance. Furthermore, improved classifi cation
performance was achieved by training separate classifi ers strati ed by gender. Classifi cation
results showed excellent accuracy, 93.1% for classifying squat quality as "poor" or "good"
and 95.3% for differentiating between high and no risk of injury. | en |
