PSI - Issue 24

Dario Vangi et al. / Procedia Structural Integrity 24 (2019) 423–436 D. Vangi et al. / Structural Integrity Procedia 00 (2019) 000–000

429

7

Fig. 6. HiL functioning scheme of an adaptive ADAS.

impact phase reconstruction process is highly time-consuming, while the time for scanning the scenario is typically close to 0.1 s. The time for scanning chiefly depends on the type of sensors and how they are mutually integrated (sensor fusion). Therefore, with a view to an onboard implementation, a database filled with the results of simulations can be conveniently considered, to be embedded into the electronic control unit ruling over the adaptive logic: relying on the database, the adaptive ADAS can easily extract the best manoeuvre associated with a critical scenario; in this case, the time required to identify the best manoeuvre corresponds to the sole time of access to the database (some milliseconds). Alternatively, the database could be replaced by appropriate solutions to minimize calculation time, like neural network systems or fuzzy logic approach (e.g., based on qualitative mechanics observations as proposed by Han (2018)). The functioning of the hardware system which includes the IR-based criteria can be outlined as in Figure 6. In a specific instant, the ADAS determines by sensors the position and heading of the opponent vehicle, its velocity, and its dimensions. The ADAS identifies the outcomes corresponding to di ff erent possible manoeuvres, calculating IR for the occupants (or retrieving it from the database) and identifying the best intervention. The information is then converted in a steering angle and a braking level to be applied. Next, the system reiterates the process: the position of the opponent and its velocity identified by sensors di ff er from the ones at the previous time step, defining therefore a new critical scenario that the system should detect and optimally handle. To highlight the potential of an adaptive logic of intervention in terms of road safety enhancement, the development of an appropriate virtual environment for the SiL implementation is required to simulate the process in Figure 3. As reported in the previous Section, the fundamental part of the cycle is the outcome retrieval associated with the single activation: with a view to an HiL implementation onboard the vehicle, by a RODM software described by the authors in Vangi et al. (2019b) a database including the outcomes associated with each braking and steering manoeuvre has been compiled, for many critical scenarios. The RODM simulation software, developed in LabVIEW, is based on the discretization of the sole vehicle perimeter through 2D beam elements, with constitutive equations similar to the ones employed in FEM models; sti ff ness of di ff erent vehicle sections is derived based on the deformation sustained in di ff erent crash test configurations as described in Vangi and Begani (2013) and Vangi et al. (2019c). The friction ellipse model described by Brach and Brach (2011) is employed for tire-road contact, governing the motion dynamics of the 3. Application of the adaptive logic