(Twizy) Fine-tune CARLA X Autoware

[dtonda8]

We want to prepare a demo for industry night. We have previously set up CARLA (a simulator) to work with an older version of Autoware that publishes /control messages to steer/accelerate the Twizy. Here's a short clip. We actually don't need to use CARLA since autoware should already publish ros messages.

As you can see in the vid above , it doesn't complete the green path but stops at the pedestrian (Red region). This is quite annoying since we want to show the Twizy turning corners and such. Here are the tasks ranked in descending priority:

• [ ] Remove pedestrians or make pedestrian locations more controlled
• [ ] Different maps (preferably the ones we created)
• [ ] (optional) Research and implement improvements to CARLA performance (currently 8 fps on Rviz)

All tasks are related to the CARLA simulator, so we'll test this with the Twizy once you've completed the tasks.

Resources: Our Carla & Autoware repo, which is based of these instructions Autoware maps Twizy's Autoware repo Links above are probably not enough, you'll need to do some more research

[dtonda8]

It seems like ~/carlaxautoware directory is broken. I remade the carlaxautoware in the ~/dbat directory. You can test it out with these commands:

First terminal:

cd ~/dbat/carlaxautoware/carla-0.9.15
./CarlaUE4.sh -RenderOffScreen

Second terminal:

export OP_AGENT_ROOT=~/dbat/carlaxautoware/carla-0.9.15/op_agent
export OP_BRIDGE_ROOT=~/dbat/carlaxautoware/carla-0.9.15/streetdrone-hil
export AUTOWARE_ROOT=~/dbat/carlaxautoware/autoware
export CARLA_ROOT=~/dbat/carlaxautoware/carla-0.9.15
export SCENARIO_RUNNER_ROOT=~/dbat/carlaxautoware/carla-0.9.15/scenario_runner
export LEADERBOARD_ROOT=~/dbat/carlaxautoware/carla-0.9.15/streetdrone-hil/leaderboard
export TEAM_CODE_ROOT=~/dbat/carlaxautoware/carla-0.9.15/op_agent
export PYTHONPATH=$PYTHONPATH:${CARLA_ROOT}/PythonAPI
export PYTHONPATH=$PYTHONPATH:${CARLA_ROOT}/PythonAPI/util
export PYTHONPATH=$PYTHONPATH:${CARLA_ROOT}/PythonAPI/carla
export PYTHONPATH=$PYTHONPATH:${CARLA_ROOT}/PythonAPI/carla/agents
export PYTHONPATH=$PYTHONPATH:${CARLA_ROOT}/PythonAPI/carla/dist/carla-0.9.15-py3.10-linux-x86_64.egg

cd ~/dbat/carlaxautoware/carla-0.9.15/streetdrone-hil/op_scripts
./run_exploration_mode_ros2.sh

After running the second commands it takes about 2 mins to load (quite slow).

---

Update

Don't worry about getting carla to work with autoware, just try to get a decent frame rate with Autoware. At the end of the day we want Autoware to publish ROS messages (check with ros2 topic list)

[dtonda8]

Closing this issue since Autoware should suffice for now, Source installation of our autoware (autoware.twizy) can be found in ~/autoware.twizy on the Alienware