Advanced methods for human movement analysis in biomechanics based on functional algorithms and multibody modelling

The human movement analysis is one of the most relevant topics of biomechanics, to which many booming research works deal with, including several areas of application, e.g. from the medical/clinics-based fields to the sport, the ergonomic and the robotic ones. More in detail, it is possible to measure the kinematic, the dynamic and the electromyographic variables that involve the complex musculoskeletal system, whichever is the analyzed motion. Regarding to the joints kinematics, which is fundamental for a quantitative study of human movement for an accurate, as well valid analysis of motion patterns, it can be noticed that these latter are influenced from the estimation of joints parameters, such as centers and axes of rotation, once both geometrical and motion models characterizing the joint connecting adjacent body segments have been assumed. The research dissertation explores this field, focusing on the so-called functional estimation of joints parameters, which is currently an open issue especially in field of gait analysis, compared to the commonly used standard approach, known as predictive method. In particular, the aim of this research is to propose a different methodology for the computation of lower limbs joints kinematics, properly based on functional algorithms, compared to the predictive approach commonly used in the laboratories of human movement analysis, so as to provide a subject-specific as well as a motion-specific analysis, considering these latter features as main challenges for currently adopted methods of movement reconstruction. One of the primary goals is to provide a simultaneous estimation of lower limbs joints parameters, based on motion data gathered and associated to the body segments. Then, joints degrees of freedom can be computed by means of specific techniques used to define the anatomical joint coordinate systems as well as the related articular conventions. The use of joint mechanical analogues of well-known kinematics, together with test rig-based measurements performed on a lower limb dummy, is firstly considered in order to evaluate the feasibility of application of functional methods with a different joint model, by means of stereophotogrammetric markers-based acquisitions. Later on, the design of customized musculoskeletal models, together with motion simulations based on real experimental data, are used to define a new functional measurements protocol during a specific motion task, such as during walking. This allows to validate the proposed functional-based procedure for joints parameters estimation, by analyzing experimental tests performed with test subjects. At this stage, also the huge problem of soft tissue artifact is faced up by means of optimization techniques allowing to compensate it, as well to proceed with the following joints kinematics computation. Results are provided in terms of accuracy, precision and repeatability of joints kinematics gait trends, often studied and analyzed at the reference laboratory of gait analysis, by comparing and discussing gait curves obtained with both the proposed functional methodology and those derived with the standard predictive approach. Secondly, taking into account the availability of body-worn inertial sensors as a valid and advantageous alternative motion capture systems of acquisition, compared to the optical ones, the feasibility of integration between these latter and functional algorithms is also evaluated to reconstruct human motion, focusing on joint kinematics. A different analytical approach as well as data post-processing techniques are considered, taking into account that available motion data rely on accelerometers, gyroscopes and magnetic sensors, instead on spatial-temporal markers trajectories computed with a video-cameras based system. A first experimental session for lower limb joints parameters estimation, as well for the related joint kinematics, is provided and graphically validated with a pilot test subject, during generic motion tasks. In particular, a data-fusion based procedure is exploited by using both inertial sensors and some markers placed above these latter, one per unit, so as to acquire also the spatial-temporal position data useful for human movement reconstruction. Thus, also a video-cameras based system is used, allowing to both record the motion itself and to provide the position information of each inertial sensor with respect to the laboratory coordinate system. The computed lower limb joints parameters are estimated with both a data-fusion based functional method using sensors data and markers spatial-temporal trajectories, and with a functional approach based only on inertial sensors data. The feasibility results are pointed out and discussed by also exploiting the graphical-based validation method. Lastly, joint kinematics results are provided by applying the proposed functional and sensors-based methodology during gait tests.