Open topic with navigation

Qualisys Biomechanics Engine

Extraction of information from session.xml

The analysis is done by the Qualisys Biomechanics Engine (QBE) and starts by extracting information from the metadata fields such as the subject information, the session information and the names of the files used for the analysis. The speed of each trial given by treadmill is transformed to calculate the running pace, i.e. time per kilometer.

Definition of the biomechanical model using the static trial

The static trial is imported and the program keeps only the frames where all markers are visible. The global coordinate system is based on the one defined during the calibration in QTM. The orientation is the same but the X, Y and Z axes corresponds now to the mediolateral (pointing to the right), longitudinal (pointing forwards) and vertical axes (pointing upwards), respectively. The biomechanical model includes the following 16 segments: pelvis (1), thighs (2), shanks (2), feet (2), thorax (1), shoulders (1), head (1), arms (2), forearms (2) and hands (2). The first step is to define the joint centers using the 35 markers.

Joint centers

The hip joint center (HIP) is calculated from the R_ASIS, L_ASIS and SACR markers using the predictive equation provided by Bell et al. (1989) and information found in Tranberg (2010). The knee joint center (KNEE) is calculated using the first running trial and a functional method called SCoRE (Ehrig et al., 2007). The SCoRE method uses the markers located on the thigh and the shank. Once the axis of rotation is found, the KNEE joint center is located at a predicted distance from the FLE marker using the work of Drillis et al. (1966) and Mukhopadhyay et al. (2010). The HIP joint center is calculated and used as virtual marker for the calculation of the KNEE joint center. The ankle joint center (ANKLE) is calculated from the FLE, FAL and TTC markers using a predictive equation based on the work of Tranberg (2010) and Gardner and Chief (1969). A local coordinate system is defined using the R_ASIS, L_ASIS, SME and TV2 markers. The shoulder joint center (SHOULDER) is then calculated as an offset from the SAE marker defined in this local coordinate system. The elbow joint center (ELBOW) is defined as the midpoint between the HLE and HME markers. The wrist joint center (WRIST) is defined as the midpoint the RSP and USP markers.

Virtual markers

The HIP, KNEE, ANKLE and SHOULDER joint centers are also used as virtual markers in the running trials. Consequently, these virtual markers are added to the list of 35 markers.

Assigning markers to the segments

Each marker is assigned to one or two segments (Table 1). Two segments are created for the upper part of the trunk. The shoulder segment is used to better assess the axial rotation of the shoulder when running.

Table 1. Segment and marker names

Segment coordinate system and marker local coordinates

A coordinate system is associated to each segment of the biomechanical model following the recommendations of the International Society of Biomechanics (Wu et al., 2002; Wu et al., 2005) when possible. For the thigh and shank segments, the SCoRE functional method used previously to determine the KNEE joint center is also corresponding to the flexion-extension axis of rotation. This axis is used in the definition of both the thigh and shank coordinate systems. The foot segment is defined by the heel and second metatarsal markers with the anterior axis defined by a virtual marker placed vertically above the second metatarsal marker in the global coordinate system. Once the segment definition is performed, the marker coordinates are expressed in their corresponding segment coordinate system. Finally, the average local position of these markers is calculated. At this point, each segment of the biomechanical model incorporates its corresponding marker local coordinates (model-based markers) and is considered as a rigid body with six degrees of freedom.

Forward kinematic function

The forward kinematic function is used to drive the biomechanical model. This function is a system of equations that gives the model-based marker coordinates in the global coordinate system using the local marker coordinates and the values of the degrees of freedom (Fohanno et al., 2014). The Cardan sequence was chosen following the recommendations of the International Society of Biomechanics.

Calculations during each running trial

The calculations start at the first frame where all markers are visible with the creation of the same virtual markers presented previously. The small gaps, length lower than 0.1s, in the marker trajectories are filled using a third order polynomial interpolation. The larger gaps in the marker trajectories are kept and an appropriate solution described below is applied. The marker trajectories are finally filtered using a 2nd order Butterworth filter with a cutoff frequency of 20 Hz.

Calculation of the joint angles and the biomechanical model position

Once the running trial is imported, the joint angles are calculated using an extended Kalman filter algorithm. This algorithm is chosen because of its ability to deal with gaps in trajectories and give smoothed joint angles while keeping a good accuracy (Fohanno et al., 2014). The basic idea is to use the forward kinematic function to drive the biomechanical model minimizing the distance between the recorded markers and the model-based markers. For the first frame, the pose of the biomechanical model is initiated with a global optimization algorithm (Fohanno et al., 2014). This algorithm minimizes the same objective function defined previously but in a least-square sense. This algorithm is not used for the remaining instants because, contrarily to the extended Kalman filter algorithm, the computational time and the noise in the trajectories would have been larger. After this first frame, the number of used degrees of freedom is decreased for some joints (Table 2). The value of the unused degrees of freedom are kept fixed during the running trial and determined using the static trial. The position of the model-based markers are stored and they will be used for the next calculations to be sure that no gaps are present in the trajectories, contrarily to the recorded markers.

Table 2. Degrees of freedom for all segments with respect to its parent. ‘YES’ indicates that the degree of freedom is used to determine the pose of the biomechanical model. ‘NO’ indicates that the degree of freedom is not used and kept at a constant value during the running trial

Elbow trajectory in the frontal and sagittal planes

The sagittal and frontal wrist trajectories are calculated using the position of the WRIST marker in the sagittal and frontal plane, respectively.

Event detection

Three types of event are determined for each side: foot strike, mid-stance and take-off.

• Foot strike (FS) is defined at local minima of the vertical mid foot velocity (FVA method of O'Connor, 2007). A time window is used.
• Mid-stance (MS) is defined as the first frame when mid foot is behind the pelvis (longitudinal axis pointing forward). Mid foot is defined as midpoint of calcaneus as defined by skeleton solver and Forefoot2 marker as defined by skeleton solver.
• Toe-off (TO) is defined as local maxima in the FM2 marker vertical acceleration (FCA method of Maiwald, 2009). A time window is used.

The time-series and trajectory patterns are normalized from foot strike (0%) to the next foot strike (100%). The right foot strike events are used for right patterns whereas the left foot strike events are used for left patterns. For unilateral patterns such as pelvis vertical position and trunk, shoulder and pelvis angles, the normalization was performed using right foot strike events. For the angle of the foot with respect to the global coordinate system, the normalization focuses on the first 25% of the contact phase. The contact phase is going from foot strike to take-off.

References

Bell, A. L., Brand, R. A., & Pedersen, D. R. (1989). Prediction of hip joint centre location from external landmarks. Human Movement Science, 8(1), 3-16

Drillis, R., Contini, R., & Maurice Bluestein, M. (1964). Body segment parameters

Ehrig, R. M., Taylor, W. R., Duda, G. N., & Heller, M. O. (2007). A survey of formal methods for determining functional joint axes. Journal of biomechanics, 40(10), 2150-2157

Fohanno, V., Begon, M., Lacouture, P., & Colloud, F. (2014). Estimating joint kinematics of a whole body chain model with closed-loop constraints. Multibody System Dynamics, 31(4), 433-449

Gardner, H. F., & Chief, O. S. (1969). A method for location of prosthetic and orthotic knee joints. Artificial limbs, 13, 31-35

Grood, E.S., & Suntay, W.J. (1983). A joint coordinate system for the clinical description of three-dimensional motions: Application to the knee. Journal of Biomechanical Engineering, 105, 136-144.

Maiwald, C., Sterzing, T., Mayer, T. A., & Milani, T. L. (2009). Detecting foot-to-ground contact from kinematic data in running. Footwear Science, 1(2), 111-118

Mukhopadhyay, P. P., Ghosh, T. K., Dan, U., & Biswas, S. (2010). Correlation between maximum femoral length and epicondylar breadth and its application in stature estimation: A population specific study in Indian Bengali Males

O’Connor, C. M., Thorpe, S. K., O’Malle, M. J. (2007): Automatic detection of gait events using kinematic data. Gait Posture 25(3), 469-74.

Tranberg, R. (2010). Analysis of body motions based on optical markers. Accuracy, error analysis and clinical applications. Doctoral Theses. http://hdl.handle.net/2077/22941

Wu, G., Siegler, S., Allard, P., Kirtley, C., Leardini, A., Rosenbaum, D., ... & Stokes, I. (2002). ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion—part I: ankle, hip, and spine. Journal of biomechanics, 35(4), 543-548.

Wu, G., Van der Helm, F. C., Veeger, H. D., Makhsous, M., Van Roy, P., Anglin, C., ... & Buchholz, B. (2005). ISB recommendation on definitions of joint coordinate systems of various joints for the reporting of human joint motion—Part II: shoulder, elbow, wrist and hand. Journal of biomechanics, 38(5), 981-992.