Montmorency Forest Wintertime Dataset
This page is dedicated to sharing our dataset for our paper entitled "Kilometer-scale autonomous navigation in subarctic forests: challenges and lessons learned".
The dataset consists of a .bag file for every run documented in the paper.
👉 Download data here (zip file of 128 GB)
Two versions of the dataset are present. "with_mapping" includes the localization mapping that was computed in real time and "no_mapping" includes the same data, with mapping filtered out. Table 2 in the Field Report paper documents all runs.
We have also added the data for runs that have been conducted in the month of October. The following link can be used to access the data for these runs.
👉 Download october data here (zip file of 22 GB)
The following table documents the runs for the latter october dataset
| Run ID | Path | Start POI | End POI | Success |
|---|---|---|---|---|
| 1 | A | Garage | Cabin | ✔️ |
| 2 | B | Cabin | Garage | ❌ |
| 3 | C | Garage | Garage | ✔️ |
In order to save space, we have only recorded the lidar packets and not the full scans. Thus, the lidar driver is required to re-compute the lidar point clouds in post-processing. Please follow the following steps:
- Add the Robosense lidar drivers in your catkin workspace;
- Add our Norlab Rosbag play package to your catkin workspace;
Then use the following command:
roslaunch norlab_bag_player warthog.launch bagfile:="<PATH_TO_BAG>"
You can also use the "rate" and "start" arguments to adjust the playback rate and start time.