-
Notifications
You must be signed in to change notification settings - Fork 196
vrx_2022 phase2_practice
The objective of this tutorial is to provide step-by-step guidance for local testing in preparation for Phase 2 - Dress Rehearsal. As described in the VRX Competition Documents, each of the VRX tasks will be evaluated over multiple trials, where each trial presents the task in a different configuration (e.g., a different waypoint location) and with different environmental conditions (wind, waves, lighting, etc.) The specifications of the environmental envelope (the range of possible environmental parameters) is included in VRX Technical Guide and the specifications of each task are in the VRX Competition and Task Descriptions - both documents are available on the VRX Website
The VRX Tasks: Examples wiki provides general instructions and example of a single trial for each task, which is a great place to start. However, if your solution is going to do well in the competition, it should be able to perform over a wide range of task and environment conditions. The purpose of this tutorial is to show you...
- How to locally test each task with multiple trials
- How to evaluate and verify your performance
For this tutorial, all of your testing will be done locally, on the host machine or a local docker container. For the actual competition, these solutions will be evaluated automatically. If you are interested in testing that aspect of the evaluation, you can setup your own evaluation setup, equivalent to the one used in the competition, using the tools described in the vrx-docker repository.
We have generated three trials for each task that cover much of the allowable task and environment parameters. These trials are notionally
- Easy - simplified task in negligible wave, wind and visual (fog) environmental factors.
- Medium - moderate task difficulty and environmental influence
- Hard - at or close to the limit of task difficulty and environmental factors.
Our intention is to execute the evaluation of submissions to the Phase 2 challenge in using very similar (but not exactly the same) trials of each task. Each trial consists of worlds and models to define the instance of the task and the operating environment.
All examples are stored in the vrx/vrx_gazebo/worlds/2022_practice directory.
You should be able to run the individual examples as follows
WORLD=stationkeeping0.world
roslaunch vrx_gazebo vrx.launch verbose:=true \
paused:=false \
wamv_locked:=true \
world:=${HOME}/vrx_ws/src/vrx/vrx_gazebo/worlds/2022_practice/${WORLD}
where you will want to change the value of the WORLD variable to match each of the worlds you want to run. Also, note that the world files (e.g., stationkeeping0.world) are in the VRX workspace. In the example above, the workspace is ${HOME}/vrx_ws. Yours might be different, e.g., ${HOME}/catkin_ws, in which case you will need to make that change.