Driving Simulation In Virtual Reality For Work Zone Design On Highway: A Validation Study

Interferences between work zones and vehicle flow can lead to a significant inc rease in accident probability. It is agreed that improving work zone design is necessary in order to provide optimal driver guidance. This paper reports on the outcomes of a survey aimed at measuring speeds on a highway in the proximity of a work zone; it, then, reports on the results of the successive experimental campaign conducted with a driving simulator rebuilding the real scenario in virtual reality. The research activities were aimed at the calibration and validation of the virtual reality system for the simulation of travelling conditions near to work zones. The surveyed work zone was located on highway A1, within Orvieto and Fabro, where a roadway was shut and the traffic diverted to the opposite roadway, according to the scheme of the Italian technical rules. Speed measurements were conducted with a laser speedmeter in the transition area, the activity area and the termination area, while in the advance warning area speeds were shot with a camera from an overpass. The images have been subsequently processed with a special software. The virtual reality experiment was developed as follows: first, the real scenario for the driving simulator was reconstructed; next, a homogenous sample of drivers was selected; eventually, the driving tests with the simulator were carried out. The field measurements highlighted that the speeds in all areas of the work zone are remarkably higher than those prescribed by the road signs. The speeds decrease below the limit only when drivers meet the physical constraint of the opening (which is only 30 m wide) for the change of roadway which compels them to adopt lower speeds. The analysis of speeds of free vehicles in the various areas of the work zone, both in real situation and in virtual reality, led to the following results: - The differences between the field speeds and the speeds from virtual reality vary from –-7.6 km/h to + 5.6 km/h. In the activity area (travelling on a single road lane with opposite flow on the other lane) the percentage deviation is + 6,1%. The mean percentage deviation in the advance warning area is – 4,5%; - the differences between the speeds observed in the real situation and those measured with the simulator resulted to be non-statistically significant for each measurement site. The bilateral Z test for non-matched samples, at the level of significance of 5%, led to accept the null hypothesis H0: the field speeds and those from the simulation belong to the same population. That means that the simulator yields the same speeds as those observed in the real situations. The values assumed by the strength of the test highlight a very low probability to run into an error of type II in the advance area and in the activity area, higher in the transition and termination areas. All these considerations enable us to validate the driving simulator of the Interuniversitary Research Center for Road Safety (CRISS) as a reliable tool for the analysis of the speeds adopted by the drivers on a highway, in several areas of a work zone.