carla/Docs/benchmark_metrics.md

64 lines
2.0 KiB
Markdown
Raw Normal View History

2018-04-10 01:42:06 +08:00
Driving Benchmark Performance Metrics
2018-04-10 01:27:15 +08:00
------------------------------
The benchmark module provides the following performance metrics, which
are related to infraction:
* Off Road Intersection: The number of times the agent goes out of the road.
The intersection is only counted if the area of the vehicle outside
of the road is bigger than a *threshold*.
* Other Lane Intersection: The number of times the agent goes to the other
lane. The intersection is only counted if the area of the vehicle on the
other lane is bigger than a *threshold*.
* Vehicle Collisions: The number of collisions with vehicles that have
an impact bigger than a *threshold*.
* Pedestrian Collisions: The number of collisions with pedestrians
2018-04-10 01:42:06 +08:00
that have an impact bigger than a *threshold*.
* General Collisions: The number of collisions with all other
2018-04-10 02:35:35 +08:00
objects with an impact bigger than a *threshold*.
These results can be computed with the metrics module, by using the following
2018-04-10 02:35:35 +08:00
code excerpt:
2018-04-10 02:35:35 +08:00
metrics_object = Metrics(metrics_parameters)
summary_dictionary = metrics_object.compute(path_to_execution_log)
2018-04-10 02:35:35 +08:00
The class is instanced with a dictionary of parameters.
These parameters could be changed by changing
The function receives the full path to the execution log and a dictionary with
parameters. It returns a dictionary with the metrics.
The parameters are:
* Threshold: The threshold used by the metrics.
* Frames Recount: After making the infraction, set the number
of frames that the agent needs to keep doing the infraction for
it to be counted as another infraction.
*Frames Skip: It is related to the number of frames that are
skipped after a collision or a intersection starts.
2018-04-10 01:42:06 +08:00
On your experiment suite class definition you could also
redefine the metrics experiment.
####Benchmark Execution
During the execution the benchmark module stores
the [measurements](measurements.md) and
[controls](measurements.md) for every single step.
These results are stored on the *_benchmarks_results*
folder.