createVehTypeDistributions.py

Creates a vehicle type distribution by sampling from configurable value distributions for the desired vType-parameters.

Example use

 <SUMO_HOME>/tools/createVehTypeDistributions.py config.txt

The only required parameter is the configuration file in the format shown below (example config.txt):

 tau; normal(0.8,0.1)
 sigma; normal(0.5,0.2)
 length; normal(4.9,0.2); [3.5,5.5]
 class; passenger
 carFollowModel; Krauss

In the config file, one line is used per vehicle type attribute. The syntax is: nameOfAttribute; valueOfAttribute [; limits]

ValueOfAttribute can be a string, a scalar value or a distribution definition. Available distributions and its syntax are: "normal(mu,sd)" with mu and sd being floating numbers: Normal distribution with mean mu and standard deviation sd. "uniform(a,b)" with limits a and b being floating numbers: Uniform distribution between a and b. "gamma(alpha,beta)" with parameters alpha and beta: Gamma distribution.

Limits are optional and defined as the allowed interval: e.g. "[0,1]" or "[3.5,5.0]". By default, no negative values are accepted but may be enabled by setting a negative lower limit.

Additional options:

• --output-file configures the name of the output file to be written
• --name Name of the created distribution
• --size Number of vTypes to be sampled for filling the distribution
• --seed Set the seed for the random number generator

Retrieving parameters from measurements of individual vehicles

To obtain mean and deviation a number of values must be obtained from the data set. The following is recommenced:

• accel: the maximum (or high percentile) acceleration for each vehicle
• deccel: the maximum (or high percentile) deceleration for each vehicle
• speedFactor: the maximum (or high percentile) quotient of speed/speedLimit for each vehicle