Closed nkoenig closed 3 years ago
Wouldn't this implementation allow "exploitation" of the simulation by simplifying detection of tile edges? Until now, it's been difficult once the robots left direct comms with teambase (with direct comms, tile edges could be estimated by large drops in rssi).
What about "smoothing" the wind values by some gradual muxing function with neighbor tile wind speeds?
Wouldn't this implementation allow "exploitation" of the simulation by simplifying detection of tile edges? Until now, it's been difficult once the robots left direct comms with teambase (with direct comms, tile edges could be estimated by large drops in rssi).
What about "smoothing" the wind values by some gradual muxing function with neighbor tile wind speeds?
Interesting. I hadn't thought about that. I'll take another look. Thanks for the suggestion.
The
fortress
branch is based on Ignition Fortress. You will need to the melodic-fortress ros_ign branch, and setIGNITION_VERSION=fortress
when running catkin_make. For exampleIGNITION_VERSION=fortress catkin_make_install
.This PR has copied the
finals_qual
world to afinals_qual_demo
world, which will be used to demonstrate features added to SubT after the final event.The first such demonstration is a local wind model. The
LocalWindPlugin
allows a world to set the wind magnitude on a per-tile basis. I've added local wind to a few cave tiles located near the staging area. Specifically, tiles 28, 29, 37, and 38.It's easiest to see the effect by increasing the magnitude to a value of 8 or higher, then move a robot into a cave tile. The wind will push the robot.