J21 Using inertial measurement units to estimate ground reaction force and knee angular velocity during decelerations


  • Luke Hitchens Cardiff Metropolitan University
  • Jasper Verheul Cardiff Metropolitan University
  • Bailey Clifford-Pryer Cardiff Metropolitan University




Despite undergoing anterior cruciate ligament (ACL) reconstruction, athletes continue to present altered ground reaction force (GRF) and knee angular velocities (AV) for 6-24 months as they return to highly demanding tasks such as rapid decelerations (Schmitt et al., 2015, Med Sci Sports Ex, 47, 1426). In-field identification of GRF and knee AV during such deceleration tasks is challenging due to the laboratory restrictions of force platforms and 3D motion capture. Therefore, the use of inertial measurement units (IMU) has been suggested as an appropriate alternative to estimate such variables in the field (Pratt & Sigward, 2018. Sensors, 3460). The study aimed to establish whether IMU-derived measures can be used to estimate GRF and knee AV that are indicative of ACL injury risk during decelerations. Following ethical approval from the Cardiff Metropolitan Ethics Committee, ten male team-sport athletes performed five maximal decelerations at three approach speeds (100, 85, and 70% relative to maximal attainable speed), for a total of fifteen trials per athlete. GRF and kinematics were measured using four Kistler force platforms and Theia3D markerless motion capture. Pearson’s correlation coefficient was used to determine the relationship between IMeasureU Blue Trident IMUs mounted at the thigh or shank and vertical GRF, horizontal GRF, and knee AV during the first step of deceleration. A measure was considered field-viable if a very large significant correlation (r ≥ 0.7; P ≤ 0.05) was observed. At approach speeds of 85 and 70%, shank AV had a very large significant correlation (r = 0.79, 0.76; P = 0.001, 0.01, respectively) to vertical GRF, a moderate to large significant correlation to horizontal GRF (r = 0.46, 0.51; P = 0.04, 0.02, respectively), and a moderate significant correlation to knee AV (r =0.46, 0.40; P = 0.004, 0.02, respectively). At 100% approach speeds, shank AV had no significant correlation to GRF (r <0.3; P>0.05) and a moderate correlation to knee AV (r = 0.40; P = 0.02). This study was the first to find a strong association between shank AV and vertical GRF for whole-body decelerations at 85 and 70% approach speeds. This supports the use of IMUs in sport-specific settings (e.g., football pitches) to better quantify vertical GRF deficits during decelerations. This can further enhance in-field monitoring of GRF deficits and better assess athletes' readiness to return to sport following ACL reconstruction. However, alternative methods are required to accurately estimate horizontal GRFs and knee AV in the field.