VR Supports the Process of Collecting Data on People’s Behavior Just before an Accident at Work Occurs
Introduction
The basic assumption of using numerical methods for the reconstruction of accidents is to aim at reconstructing the real situation in which the accident occurred by means of computer simulation. Thanks to the numerical modeling of physical phenomena, it is possible to reconstruct the real course of events, including its effects and causes, based on the laws of physics. To simulate accidents, numerical models of the human body are most often used, which faithfully reproduce the kinematics of the human body and enable the assessment of injuries. At present, passive computer models of the human body are used for the computer reconstruction and simulation of fall-related accidents, that is, those that maintain general kinematics but do not take into account body movements resulting from muscle tension. This means that during the simulation of a fall, the computer model of a human body falls to the ground inertly. An important factor affecting the trajectory of a falling human model is, however, its response when the loss of balance occurs. At that time, a human determines the baseline conditions that are the input data used to start the numerical simulation of the fall. Similar problems were noticed as a result of analyzing road accidents. It turns out that a human’s response during a car accident also has a very significant impact on the trajectory of movement for the injured person. Therefore, a need to develop so-called active human models appeared, that is, models that enable the simulation of movements resulting from muscle tension. One of the most advanced is the active human model developed by TASS-Safe, which enables the simulation of reactions during a rollover (Meijer et al. 2008). The model developed by this team is made out of the Net-torque technique, meaning that it is activated through the help of “virtual servomotors”, generating the appropriate moment for rotation in the simulation of joints and individual spine vertebrae. To develop this model, it was necessary to conduct research with the participation of volunteers, which enables the typical reactions of a person involved in this type of accident to be determined (Muggenthaler et al. 2008).
Active models of individual body parts are also created for the needs of the automotive industry. An example could be the model of the upper limb which is controlled by numerical muscles (Budziszewski et al. 2008; Ósth et al. 2010). Another example is the head and neck model developed by the University of Loughborough, which is controlled by numerical muscles and used to study whiplash, that is, neck injuries caused by a sudden, rapid back-and-forth movement that are most often caused as the result of a rear impact (van Lopik and Acar 2004).
To develop such models, it is necessary to conduct research with the participation of volunteers in order to determine human responses to certain types of external stimuli. Research on human’s responses is conducted not only for the needs of creating such numerical models, but also for the needs of medicine. Immersive VR techniques are increasingly used for this type of research. This technique is used everywhere where testing in real conditions would either be too dangerous or a multitude of scenarios w'ould make testing in laboratory conditions impossible. This technique has been successfully used for research into the treatment of acrophobia (fear of heights) (Rothbaum et al. 1995a,b; Krijn et al. 2004; Bush 2008; Carmen Juan and Pérez López 2010). To check whether this type of research is a reliable reflection of real conditions, an experiment was carried out that compared the feelings of subjects placed at a certain height above the ground and was then repeated, with the difference being that the height above the ground was presented in the form of a synthetic environment using an HMD (in reality, in this scenario the subject was standing at a height of 0 m; Clew'orth et al. 2012). The level of fear and sense of presence was admittedly higher in the real conditions than when using the VR techniques. Nevertheless, the obtained results did confirm that the subjects tested using VR had the impression that they were at a certain height above the ground.
Research is also conducted on maintaining balance in a simulated environment for the needs of neurology (Tossavainen et al. 2003; Lee et al. 2004; Horlings et al. 2009). One of the more interesting studies was conducted by a team led by Hsiao-Yu Lee. The research consisted of presenting an image of a room that at one point tilted to simulate the rotation of a human head. These studies showed that some subjects reacted to the image change by tensing their muscles and even changing their body position. In the next test scenario, the platform on which the subject stood was also tilted. In this case, the results showed a significant increase in muscle tension. This configuration of tests meant that the subjects were successfully thrown off balance.
With the aim of broadening knowledge about the causes and dynamics of falls from heights, as well as developing a numerical human model that enables human responses in the moment of a loss of balance and in the initial phase of the fall to be considered, the CIOP-PIB team conducted research with the participation of volunteers as well as using VR techniques and vision-based motion-capture systems (hence, contrary to the active human model developed by TASS-Safe, referred to earlier, this model therefore applies to other types of accidents at work). The use of VR allowed the subject to be “immersed” in a virtual environment that presented a view from a height of a few or several meters above the ground. As a result, the subjects had the impression that they were at a certain height above the ground. This solution enables conducting safe research at a level of just several centimeters above the ground and also enables the use of different scenarios depending on the needs of thestudy. A tilting platform on which the test station would stand was also planned to be used, which would increase the sense of spatial presence and allow a controlled loss of balance to be induced in the subjects. The results of the tests include the trajectories of individual body parts along with the angles of bend in individual human joints.
