Our ESR14-Allessando Bonfiglio-defended his thesis successfully on April 14th!
Alessandro was the only ESR doing his PhD in an industrial setting (Euleria) and received his PhD degree from the University of Trento in Italy.
He presented his PhD entitled: “Applicacability of IMU in Clinical Research” and received questions from his committee members.
Congratulations Alessandro!!
Abstract:
Inertial Measurement Units (IMU) represent a staple in contemporary motion tracking technology, and they have countless applications in sports, animation, clinical research and biomechanics due to their portability, reliability and low cost. However, several challenges related to orientation estimation, sensor-to-segment calibration and mitigation of soft tissue artifacts prevent such innovative technology from widespread use among the general population. To fill this gap, this work investigated three common sensor-to-segment calibration techniques, namely N-pose (NP) static calibration, functional calibration (FC) and manual alignment (MA) calibration for the Humerothoracic, Elbow and Wrist joints during a variety of movements in order to shed light on each ideal use case and application. Furthermore, this manuscript presented two potential applications of IMU in clinical research to monitor the motion of the wrist and lower body kinematics.
Each calibration was compared against an active optical motion capture reference system, and performance was evaluated by computing ROM error (ε), Root Mean Squared Error (RMSE) and offset as dependent variables (DVs). Statistical differences between calibrations and interaction effects calibration × axes and calibration × tasks were analysed using a three-way repeated measures ANOVA for each joint and DV.
Results were different depending on the joint. No statistical differences were observed between NP, FC and MA for the humerothoracic joint for any DV. Therefore, we recommended adopting the NP calibration due to its simplicity and quickness. The elbow joint analysis yielded significant differences in RMSE between calibrations. Specifically, the MA calibration proved superior compared to the other options, resulting in the advised calibration for this joint. The wrist joint analysis resulted in significant differences among calibrations in most DVs, highlighting FC as the most accurate calibration technique. Therefore, FC appears to be the most suitable candidate for a wrist motion tracking system using two IMUs on the lower arm and hand, which can apply to several musculoskeletal or neurological conditions due to its unrestricted setup requirements. Finally, despite its proven disadvantages for elbow and wrist joints, NP calibration can find applications in complex multi-joint systems due to its ease of use and setup quickness. Therefore, we presented a proof of concept for a motion tracking system to monitor lower body kinematics reliant on 5 IMUs operating on an NP calibration.
This work provided practical and applicable guidelines for the use of IMU sensors in clinical practice, highlighting the advantages and disadvantages of each calibration depending on the joint, type of movement and rotation axis. These results can aid in driving innovation in clinical motion analysis by acting as a foundation for accuracy and reliability metrics on IMU, as well as identifying areas for improvement.