Skip to content

harvard-edge/AVPerf

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AV Benchmarks

Scenes

This project aims to create an AV benchmarks suite to evaluate AV system. We adopt the state-of-the-art Apollo AV stack and create three safety-critical scenes to conduct end-to-end analysis. We collect AV metrics, including collision, minimum distance, and safe end-to-end processing rate to evaluate the AV system on different scenes.

There are three safety-critical scenes listed below:

  • cut-in
  • cut-out
  • vehicle following

Please reference to SVL documentation to run Apollo AV stack with SVL simulator. For Apollo AV stack, please clone the latest version of Apollo and replace the modules folder with the apollo_modified_modules provided in this repository.

To run a single scenario

python safety-critical/$SCENE $FPS $INDEX (e.g., python safety-critical/cut-in.py 5 1)

To run multiple scenarios

cd scripts
./run.sh

Inside run.sh, user can modify FPS_start and FPS_end to change the range of sweeping FPS, and change iteration for the experiment runs for each FPS.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published