• QueueOutput

QueueOutput

The idea behind this output option is to detect the queue in front of the controlled/uncontrolled junction. The queue length is calculated using the end of the last standing vehicle.

Instantiating within the Simulation#

The simulation is forced to generate this output using the option --queue-output <FILE>. <FILE> is the name of the file the output will be written to. Any other file with this name will be overwritten, the destination folder must exist. Using the option --qeue-output.period <TIME>, the times the queue lengths will be written can be restricted to the given period. Then, the other time steps will be skipped.

The speed below which a vehicle is counted as queued can be configured with the option --queue-output.speed-threshold <FLOAT> (default 1.39 m/s, i.e. 5 km/h).

By default a (possibly empty) record is written for every time step (or interval). With the option --queue-output.skip-empty time steps / intervals that contain no queue are omitted.

Data Format#

The output file format is determined by the file extension (or the option --output.format). Besides XML, the column based formats CSV (.csv) and Parquet (.parquet) are supported. Because these formats require a homogeneous set of columns, the option --queue-output.skip-empty must be set when using them (otherwise empty time steps would create rows without queue data).

Generated Output#

The generated XML file looks like this:

<queue-export>

  <data timestep="<TIME_STEP>">

  <lanes>

    <lane id="<LANE_ID>" queueing_time="<LANE_QUEUEING_TIME>" queueing_length="<LANE_QUEUEING_LENGTH>"
    queueing_length_experimental="<LANE_QUEUEING_LENGTH_EXPERIMENTAL>"/>

    ... next lane ...

  </lanes>
  </data>

  ... next timestep ...

</queue-export>
Name Type Description
time_step (simulation) seconds The time step described by the values within this timestep-element
id id The id of the lane
queueing_time seconds The total waiting time of vehicles due to a queue
queueing_length meters Thus the length from the junction until the final vehicle in line
queueing_length_experimental meters The length of the queue, thus until the last vehicle with a speed lower than 5 km/h

Mesoscopic Simulation#

When running with mesoscopic simulation, queues are measured per segment queue instead of per lane. Vehicles within a segment queue are ordered by their entry time and the queue length (in vehicles) is the position of the last queued vehicle, counted from the head of the queue. A vehicle counts as queued if the slope of its (interpolated) position trajectory lies below the speed threshold. This slope is derived from the earliest exit times chained along the queue (the same computation used by --meso-interpolate-pos) and therefore reflects blocked vehicles further ahead, whereas the plain vehicle speed in mesosim only drops once a vehicle becomes the queue head. The generated lane elements carry the additional attributes segment (index of the segment within the edge) and queueing_count (queue length in vehicles). The attribute queueing_length gives the distance from the segment end to the back of the last queued vehicle when --meso-interpolate-pos is active and the summed vehicle lengths (including minGap) up to the last queued vehicle otherwise.

Aggregated Output#

When setting the option --queue-output.aggregation <TIME>, per-timestep data is replaced by per-edge queue length statistics aggregated over the given period (typically a traffic light cycle):

<queue-export>

  <interval begin="<BEGIN>" end="<END>">
    <edge id="<EDGE_ID>" samples="<NUM_SAMPLES>"
      maxQueueLengthInVehicles="..." medianQueueLengthInVehicles="..." percentileQueueLengthInVehicles="..."
      maxQueueLengthInMeters="..." medianQueueLengthInMeters="..." percentileQueueLengthInMeters="..."/>

    ... next edge ...

  </interval>

  ... next interval ...

</queue-export>

Each lane (or meso segment queue) with at least one queued vehicle contributes one sample per recorded time step. The median and percentile values are computed over all samples of the interval with linear interpolation between order statistics (matching numpy.percentile). The reported percentile defaults to the 95th and can be configured with the option --queue-output.percentile <FLOAT> (a value between 0 and 100). Edges without any queued vehicles during the interval are omitted.

Notes#

This output option should offer some information about the queues in front of the junctions, which can be used in cases of routing or V2X communications.

Visualization example#

Queuing times over time#

Generated with plotXMLAttributes.py.