README

Contributor: Anurag Maurya

File structure:

The repository has 5 packages:

1) DR_SPAAM_ROS: Package for detecting person from LiDAR data

2) rf2o_laser_odometry: For generating fake odometry in some cases

3) rplidar_ros: Standard package for getting sensor data from RPLiDAR

4) wc_navigation: Relevant launch files for navigation of wheelchair in 2D mode.

5) wc_navigation3D: Relevant launch files for navigation of wheelchair in 3D mode.

---

Some more details regarding files:

The goal_set.py node and sock.py node use socket API to get goal positions from another computer working on BCI headset data and publish the corresponding goal to the navigation stack.

---

Launching navigation

2D Mapping

For mapping the arena using Hector SLAM

Launch rplidar node

roslaunch rplidar_ros rplidar.launch

Launching the mapping node

roslaunch wc_navigation map_hector.launch

Saving the map

rosrun map_server map_saver -f test_1

---

2D Navigation

To launch the navigation using only RViZ

roslaunch wc_navigation wc_navigation.launch

To launch the navigation and integrate the BCI headset

roslaunch wc_navigation wc_bci.launch

---

3D Mapping using Realsense

Using D415 here, other cameras might have different configurations

1. Install Realsense wrappers

GitHub - IntelRealSense/realsense-ros at ros1-legacy

sudo apt-get install ros-$ROS_DISTRO-realsense2-camera

2. Install RTAB-MAP

https://github.com/introlab/rtabmap_ros

Can install from source or apt too

sudo apt-get install ros-$ROS_DISTRO-rtabmap-ros

3. Run into terminal

roslaunch realsense2_camera rs_camera.launch align_depth:=true

Make sure that /camera/aligned_depth_to_color/image_raw is being published

4. Run into terminal

roslaunch rtabmap_ros rtabmap.launch rtabmap_args:="--delete_db_on_start --Optimizer/GravitySigma 0.3" depth_topic:=/camera/aligned_depth_to_color/image_raw rgb_topic:=/camera/color/image_raw camera_info_topic:=/camera/color/camera_info approx_sync:=false

D435i

Run into terminal

rosrun imu_filter_madgwick imu_filter_node _use_mag:=false  _publish_tf:=false _world_frame:="enu"  /imu/data_raw:=/camera/imu  /imu/data:=/rtabmap/imu

Run into terminal

roslaunch rtabmap_ros rtabmap.launch	rtabmap_args:="--delete_db_on_start --Optimizer/GravitySigma 0.3"	depth_topic:=/camera/aligned_depth_to_color/image_raw	rgb_topic:=/camera/color/image_raw	camera_info_topic:=/camera/color/camera_info	approx_sync:=false	wait_imu_to_init:=true	imu_topic:=/rtabmap/imu

Mapping using entire setup

roslaunch wc_navigation3D wc_3d_mapping.launch

Navigation in 3D using entire setup

	roslaunch wc_navigation3D wc_3D2.launch

Getting Person detections

roslaunch dr_spaam_ros dr_spaam_ros.launch

Launching Low level controller

rosrun rosserial_python serial_node.py /dev/ttyACM0

Demonstrations

2D navigation in a lab environment

3D mapping in a lab environment

3D navigation in lab environment

Videos

2D navigation

2D_nav.mp4

3D navigation

3D_nav.mp4

Issues

uvc streamer issues:

IntelRealSense/realsense-ros#2149

References

@article{Jia2021Person2DRange,
  title        = {{Self-Supervised Person Detection in 2D Range Data using a
                   Calibrated Camera}},
  author       = {Dan Jia and Mats Steinweg and Alexander Hermans and Bastian Leibe},
  booktitle    = {International Conference on Robotics and Automation (ICRA)},
  year         = {2021}
}