SIMULATED TRAFFIC FOR THE VTI DRIVING SIMULATOR: A PRELIMINARY STUDY

A driving simulator is an advanced piece of equipment used to create realistic driver sensations in a laboratory environment. The the Swedish National Road and Transport Research Institute, (VTI) driving simulator has an advanced motion system. Today the VTI driving simulator does not include any model that generates and simulates ambient traffic. The aim of this master thesis was to construct a model that generates and simulates vehicles for the VTI driving simulator. In order to achieve this, a survey of other similar models around the world was performed. It was found that there are at least ten similar models. The model is valid for a straight, dry and plane motorway without any entering and exiting lanes. The model is named Intelligent Traffic Generator, INTRAG. There are five different areas of three types in INTRAG. The three types are: simulated area, candidate area and generation area. The area closest to the driving simulator is the simulated area. In this area the vehicles are simulated according to models for driving behavior. The models used for car-following, lane changing and acceleration are based on the model Traffic Performance on Major Arterials, TPMA, and the VTI Traffic Simulation Model. INTRAG also includes models for deceleration, transformation of basic desired speed and communication between vehicles with turn and brake signals. The areas on both sides of the simulated area are the candidate areas. These areas are buffer zones of vehicles. The vehicles in these areas are seldom updated. When updated, this is not done according to any models for driving behavior. There are two generation areas in INTRAG, one at the end of each candidate area. In these areas, new vehicles are generated. Vehicles that have traveled out of the model are removed in this area. The thesis also includes the task to construct a communication link between INTRAG and the driving simulator. The Internet protocol IP and the transport protocol UDP were used. The packages that are sent between INTRAG and the driving simulator are built according to a structure that was developed. INTRAG has been implemented in the programming environment Borland Delphi 5. An application that can be used as a substitute for a real driving simulator was created to test INTRAG. The substitute has also been used during the model verification and validation. It was found, during the model verification, that the flow in average is 3.6 to 9.9% too low. The reason is probably non-valid measurement methods. Besides this, the model seems to be implemented correctly. INTRAG has not yet been tested within the VTI driving simulator. The tests that have been done with the substitute provide an indication that the model may be a good representation of the reality. (A) This report is also available electronically via Internet at http://www.vti.se/PDF/reports/N32-2003.pdf.