Invited Paper - Simulation and Modeling Methodologies, Technologies and Application - page 7

vehicle having a driver model and a vehicle dynamics model. The driver model con-
sists primarily of four parts; a psycho-physical following model [7]; [8], a lane chang-
ing model, a launch model and a speed holding model. The output of the driver model
is the driver's desired speed, acceleration and lane angle. These are later modified to
conform to the vehicle's performance limits.
In addition, VISSIM has two features that may be important in future work; the
ability to introduce a user driver model with lane-changing and following behavior,
and an interface for the dynamic routing function that allows exploration of routing
algorithms using a programming interface.

2.3

The Scenarios

Three representative road types were used to build traffic scenarios. The road types
were chosen to be exemplary of the types of roads that might populate a full scale
analysis project; the residential street, urban highway and limited access highway.
Road models were based on actual roads near Dearborn, MI, USA where the work
was done and coded into the traffic simulation program. This allowed easy access to
collect data and calibrate the models (see Fig. 2). Fig. 2 presents a 3-mile stretch
along US I-96 (between exit 179 and 183) that was coded into the simulator as a rep-
resentative section of freeway. The base model is a 3 lane road with no ramps enter-
ing or leaving the freeway. The traffic composition included 4% heavy goods vehicles
and 2% battery electric vehicles. The remaining 94% were internal combustion ve-
hicle of varying lengths consistent with personal transportation. To differentiate be-
tween a BEV and an internal combustion vehicle drivetrain, different desired accele-
ration profiles (speed vs. maximum acceleration desired by the driver) have been
used. The vehicle input was set at 5000 vehicles per hour over 3 lanes. Stochastic
distributions of driver desired speeds are defined for each vehicle type within each
traffic composition. The desired speed of both the conventional cars and the BEV was
a roughly normal distribution with a mean of 62 MPH (100 km/h) distributed between
83 MPH (130 km/h) and 50 MPH (80 km/h). At this speed aerodynamic drag exceeds
all other vehicle specific load components except possibly accessory loads. If not
hindered by other vehicles, a driver will travel at his desired speed with variations
determined by the driver following model.
The urban highway model is based on a 6 mile stretch along US-24 (Fig. 3) with
multiple traffic signals at intersections 1 mile apart. The vehicles enter at one end of
the road and exit only at the other end. There were a total of nine synchronized traffic
signals along the road, with multiple signals at some intersections. Although the ac-
tual road had 4 lanes along part of the stretch, the base model had 3 lanes throughout
to simplify interpretation of the results. The traffic was composed of 98% convention-
al cars and 2% battery electric vehicles. The desired speed varied between 42 and 48
MPH (68 km/h - 77 km/h).
Traffic light timing was on roughly 60 second cycles such that during a typical
evening rush hour packs of about 90 cars build up at a red light. The light would
change to green and the vehicles would launch from a standstill. Except for the lead
vehicles, each vehicle's launch rate was limited by the vehicle ahead. The pack of
vehicles would reach the next light and stop for a few moments, and then continue to
the next light.