Skip to content

v2

Compare
Choose a tag to compare
@bjsowa bjsowa released this 14 May 18:21
· 17 commits to master since this release

The initial release of the Navigation Task Simulation for the ERC2021 Remote Competition.

Changes

Now using Git Submodules to pull external repositories

This repository now links specific versions of external repositories with ROS packages using Git Submodules. To avoid any conflicts with the previous packages, we recommend cloning the repository again and following the new instructions from README.

New packages

To avoid confusion about packages for the stock Leo Rover and the ERC-edition one, we decided to create separate packages for Leo Rover ERC-edition. These packages are mostly similar to the counterparts from the stock Rover, but have names with leo_erc_ prefix instead of simply leo_. The leo_erc_ packages sometimes use resources from the newest leo_ packages available in the ROS distributions, so make sure your system is up-to-date.

Updated the ERC payload models

The URDF model of the robot has been updated with the ERC2021 edition of the payload for Leo Rover.

Added artificial landmarks to the terrain

The simulation environment now includes models of the artificial landmarks with ARTags, similar to the ones that will be present on the real terrain.

Added simulated visual odometry from ZED2 camera

Visual odometry is now computed from the simulated ZED2 camera to replicate the zed2/odom topic from zed-ros-wrapper node that will be run on the real robot.

Added wheel odometry pose estimates

The newest version of leo_firmware added 2D pose estimates from wheel encoders published on the wheel_pose topic and resettable through core2/reset_odometry service. We replicated this feature in the simulation.

Added Ground Truth position information

An information about the actual position of the robot in the simulation environment is now published on the ground_truth topic. It has been added to help validate the performance of localization systems, but it won't be included on the real robot (for obvious reasons).