Statistics Output

Statistics Output is another output generating tool to easily measure basic simulation outcomes. You will be able to obtain short or detailed results of the simulation, e.g. travel times or the average speeds of groups of vehicles, or the average flow on induction loops.

Configuration

• The main configuration file for all output generators is located at <scenarioName>/output/output_config.xml

In order to use the Statistics Output, the attribute enabled of the root element output must be set to “true”, as shown in the following listing.

<output id="statistics" enabled="true" loader="com.dcaiti.mosaic.fed.visualizer.StatisticsVisualizerConfig">
   <statistic>
      [...]
   </statistic>

   <subscriptions>
      <subscription id="VehicleRegistration"/>
      <subscription id="VehicleUpdates"/>
      <subscription id="TrafficDetectorUpdates"/>
   </subscriptions>
</output>

Configuration header for Statistics Output

Specification

In this section, we take a closer look at the Statistics Output by using examples and demonstrations. For each type of retrieving data we create a <statistic></statistic> block. Inside the block we define one certain data type we want to retrieve in a element. If you want to retrieve different data types, just create another <statistic> block for each of them.

You can also set the wanted file name in the attribute filename of the statistic element. If the attribute has not been given, each <statistic> block will get the name accordingly to the order number, for example 1. StatisticsVisualizer-Block.csv.

In the output attribute two options (short|verbose) can be selected. The short option provides us a compact log file with information about only the highest level of the retrieved data (e.g. aggregate values of grouped vehicles) in contrast to verbose option which also provides informations about every individual vehicle in each group.

For a successful start, the element source must be placed in the first position in the statistic children element list. Different options of possible data types, one of which must be specified in the source element can be seen below.

<statistic filename="ChooseItYourself" output="short">
	<source>NameOfSource</source>
</statistic>

Source options of Statistics Output

Application

This section will demonstrate the basic idea and usage of the Statistics Output depending on the individual requirements. Next to retrieving raw data, the Statistics Output has further features for processing of the obtained data.

1. source: Data to obtain, choose between:
   • VehicleSpeeds - Obtain the speeds of the vehicles of each simulation step.
   • VehicleStops - The total number of stops during the journey of each vehicle.
   • VehicleTravelTimes - The total travel time in s of the vehicles.
   • VehicleDelayTimes - The deviation of the travel time compared to the fastest travel time possible for the vehicles (in s).
   • VehicleTravelledDistances - The travelled distance of the vehicles in m.
   • VehicleFuelConsumptions - The fuel consumptions of the vehicles in l per km.
   • VehicleHeadways - Obtain the headway towards the leading vehicle of each vehicle for each simulation step. To obtain this value, an application has to be deployed on the vehicles which activates the front distance sensor.
   • DetectorFlow - The flows of each subscribed induction loop.

2. group-by: The vehicles will be grouped by its vehicle type name (VehicleType), group they belong to (VehicleGroup), or obtained data value (e.g. Interval:200 categorizes values into groups of size 200).

3. aggregation: Average | Harmonic aggregation of the obtained values. An attribute deviation can be set to true or false (it’s false if the attribute is not given). This attribute can be used for grouped values to get the deviation of each value from the aggregated group value or to get a standard deviation based on biased sample variance for groups (in the short output version).

4. filter: Filtering with the attribute filterType (possible values are keep and remove).

   • Filtering by required value slots with two options to specify them: MoreThan:Value or LessThan:Value (e.g. MoreThan:5 to collect values which are bigger than 5 only)
   • Filtering by vehicle type. VehicleType:Type (e.g. VehicleType:Car to collect values only of vehicles of type “Car”)
   • Filtering by time. Time:From-To (e.g. Time:0-100 to collect values only of the first 100s of simulation time)

The following example will show an example of how you can specify the Statictics Output according to your desired criteria. VehicleTravelTimes are the data we want to retrieve from vehicles and we want to group vehicles by the abstract group we can define in mapping configuration file (see e.g. Barnim scenario) and then calculate the average vehicle travel time value for each of these groups.

<output id="statistics" enabled="true" loader="com.dcaiti.mosaic.fed.visualizer.StatisticsVisualizerConfig">
	<statistic filename="AverageVehicleTravelTimes" output="short">
		<source>VehicleTravelTimes</source>
		<group-by>VehicleGroup</group-by>
		<aggregation>Average</aggregation>
	</statistic>

   <subscriptions>
      <subscription id="VehicleRegistration"/>
      <subscription id="VehicleUpdates"/>
      <subscription id="TrafficDetectorUpdates"/>
   </subscriptions>
</output>

Getting the Average time by vehicle class

You can also combine filters if you want to get a certain interval with upper and lower boundaries. With the following input instruction, only vehicles with the obtained data values between 250 and 500 will be left after filtering.

<filter filterType="keep">LessThan:500</filter>
<filter filterType="remove">LessThan:250</filter>

An example for filtering

Please notice that some sources are being not only obtained in each simulation step but also collected for further processing as separate values for each of these steps (like VehicleSpeeds, VehicleHeadways). These data types need to be aggregated to one value per vehicle if you want to group them by value or filter them.

For demonstration, the Statistics Output is configured for the scenario Barnim and calculates the average travel times of the vehicles and additionally groups them. As a result, the simulation produces the following CSV file in the log directory:

Group;Value;Total;
AdHoc;399.14;24;
Cellular;463.87;12;
Unequipped;459.18;84;

The AverageVehicleTravelTime.csv file produced by the Statistics Output in the Barnim scenario