HuPOSE : Human-like posture generation and biomechanical analysis for human figures

Over the years an increasing attention has been devoted to ergonomic analyses even from the early stage of the design process. Ergonomic and human factor evaluations often require building a physical mock-up in order to provide an assessment of discomfort and ease of use. This process, using traditional methods, is very time demanding, especially when the design has to be modified and revalidated. Digital mock-up instead, enables manu- facturers to design digital prototypes of a product in full details, simulating its functions and predicting interaction among its different components. In order to take advantage of digital simulation to conduct ergonomic assessments digital substitutes of human beings (also called digital humans), able to interact with the digital mock-up in simulation environment, are required. Since these digital humans are required to simulate human beings in digital environments their resulting movements must be as human-like as possible. Although these digital human simulation tools are now advanced enough to correctly predict human-product and human-process interaction, even before a physical prototype is constructed, the animation process is still very time demanding, mainly because it still relies on key frame techniques. Moreover, the accuracy of the resulting simulations are strongly related to the experience of the operator. The aim of this thesis has been to develop an algorithm capable of speeding up the animation process of digital humans. An algorithm capable of conducting biomechanical analyses has been developed as well. Chapter 1 provides a general introduction underlining the need to use digital human simulation tools from the early stage of the design process. The main applications of digital technologies in industrial world are presented as well. Chapter 2 provides an overview of the the main digital human simulation tools currently available, highlighting their advantages and disadvantages. Chapter 3 describes the mathematical theory underlying the developed HuPOSE model. Both the kinematic and the biomechanical model are presented. The main contribution is the formulation of the inverse kinematic problem in terms of a single CLIK algorithm, using an Augmented Jacobian matrix. This approach suggested also the possibility of computing the static torques at the joints of a digital human by means of kineto-static duality. The computation of the static torques allowed to conduct a biomechanical analysis, in reference to a load-lifting task, very easily. Chapter 4 discusses several possible application for the developed HuPOSE model. Simulation in virtual environment have been conducted using Matlab–Simulink in order to show the ease of motion planning for a human figure. The implemented whole-body motion control technique takes into account the position of the centre of pressure of the digital human. This technique allows to achieve quite natural movements in spite of the limited number of task related control points considered. A biomechanical analisys is presented as well, whose results are results are in good agreement with literature data. Chapter 5 contains the main results achieved, remarks and proposals for future developmen