Using IMUs for the assessment of knee flex-extension angle in presence of soft tissue artefacts

Overweight people commonly conduct a sedentary lifestyle because of difficulties in walking. To quantitatively monitor their gait and reduce soft tissue artefacts, optoelectronic systems are usually adopted, but also inertial measurement units (IMUs) have shown promising results. Considering overweight children, the long preparation time of optoelectronic systems could be a disadvantage. On the contrary, IMUs could simplify the walking acquisition protocol. In light of these considerations, a pilot study was conducted exploring the use of IMUs during the flex-extension of a polycentric knee model in presence and absence of soft tissue. In detail, the aims of the work were: (i) to evaluate the filtering effect on knee flex-extension movements calculated using IMUs, and (ii) to calculate knee flex-extension movements from linear acceleration and angular velocity of lower limb segments. Two silicone implants were obtained using silicone ECOFLEX 0030, which mimics soft tissue properties. The two implants had different size to fit the thight (diameter 9.5-10.5 cm, thickness 3.5 cm) and the shank (diameter 9 cm, thickness 2.5 cm) segments, respectively. The implants were fixed on the lateral part of the thigh and the frontal part of the shank of the polycentric knee model. Three knee flex-extension trials were performed: (1) thigh segment kept steady in vertical direction and shank segment moving, (2) thigh segment kept steady in horizontal direction and shank segment moving, (3) shank segment kept steady in vertical position and thigh segment moving. For each trial, two Xsens IMUs were fixed on the thigh and two on the shank: (i) one on the rigid segment and (ii) one on the implant. In addition, one Xsens sensor was fixed on the ground as reference. For comparing knee flex-extension movements in presence and absence of soft tissue, Xsens orientation data calibrated with an internal Kalman filter were used. Rotation matrices of each IMU sensor were first referred to a static position. Then, knee flex-extension angle was calculated from the relative orientation between IMUs on the shank and on the thigh. This procedure was repeated for the sensors on the rigid segments and on the implant. Results showed very similar trends in presence and absence of soft tissue. To calculate knee flex-extension movements from linear acceleration and angular velocity of lower limb segments, the raw data collected with Xsens were used. A procedure consisting of sensor fusion and a Kalman filter was implemented both for IMUs on the rigid segments and on the implant. This post-processing allowed comparing results obtained from raw data to those obtained from Xsens showing an overall agreement. In conclusion, this pilot study showed promising results; however, further investigations on overweight children should be conducted to confirm the suitability of IMUs to assess knee flex-extension angle.