UGVNavigationSystem

This is a UGV navigation system based on occupancy grid map in GPS-denied environments. We use open-source mapping algorithm Traversability Mapping to build the environment map first and use map_server ROS package to save the map to disk. we then back to the start point of mapping algorithm. based on the preliminary map model, once there is a goal point to send, the UGV will launch towards the target until it reachs the destination. We use LEGO-LOAM to locate the UGV, use ROS Navigation Stack to find optimal path and track the generated trajectory and use mqtt to communicate with remote terminals.

1. Prerequisite

1.1 first do

Ubuntu and ROS Ubuntu 64-bit 18.04 and ROS Melodic. ROS Installation Create a catkin workspace

sudo apt install ros-melodic-catkin
mkdir -p ~/catkin_ws/src
cd ~/catkin_ws/src
catkin_init_workspace
cd ~/catkin_ws/
catkin_make

echo "source ~/catkin_ws/devel/setup.bash" >>  ~/.bashrc
source ~/.bashrc

#check if success
echo $ROS_PACKAGE_PATH

ROS Package dependecies

sudo apt install ros-melodic-navigation

1.2 Receive lidar data

• install the dependencies in rslidar_sdk
• download via git

cd ~/catkin_ws/src
git clone https://github.com/RoboSense-LiDAR/rslidar_sdk.git

• compile with ROS catkin tools following rslidar_sdk
• modify lidar_type: RSHELIOS in config.yaml
• run roslaunch rslidar_sdk start.launch
• open rviz, you can see the pointcloud in frame /rslidar

1.3 LeGO-LOAM

• download via git

cd ~/catkin_ws/src
git clone https://github.com/RobustFieldAutonomyLab/LeGO-LOAM.git

• modify global parameters in utility.h

extern const string pointCloudTopic = "/rslidar_points";
//RSHELIOS
extern const int N_SCAN = 32;
extern const int Horizon_SCAN = 1800;
extern const float ang_res_x = 0.2;
extern const float ang_res_y = 1.0;
extern const float ang_bottom = 16;
extern const int groundScanInd = 7;

1.4 Mapping

• download via git

cd ~/catkin_ws/src
git clone https://github.com/TixiaoShan/traversability_mapping.git

• modify global parameters in utility.h

// RS-Helios vertical and horizontal points per scan 
extern const int N_SCAN = 32;
extern const int Horizon_SCAN = 1800;
// Robot Params
extern const float robotRadius = 0.6;
extern const float sensorHeight = 0.9;
// length of the local occupancy grid map (meter)
extern const int localMapLength = 300; 

• run the launch file roslaunch traversability_mapping online.launch
• save grid map topic:=/occupancy_map_local name:=center_map2 rosrun map_server map_saver -f center_map2 map:=/occupancy_map_local

1.5 Path planning and Control

• download this repository via git

cd ~/catkin_ws/src
git clone https://github.com/WillenChung/UGVNavigationSystem.git

• modify parameters in move_base.launch&car_ctr.launch

<!-- move_base.launch -->

<!-- load map config file -->
<arg name="map" value="new_map_c.yaml" />

<!-- car_ctr.launch -->

<!-- configure car speed 0-1000 -->
<arg name="car_speed" value="400"/>
<!-- configure PID params -->
<arg name="P" value="10.0"/>
<arg name="I" value="0.12"/>
<arg name="D" value="0"/>
<arg name="IMax" value="50"/>
<arg name="PIDMax" value="150"/>
<arg name="I_Threshold" value="15"/>
<arg name="I_DeadArea" value="1"/>
<!-- configure following path topic -->
<arg name="follow_path_set" value="/move_base/NavfnROS/plan"/>

1.6 Communication with remote

2. Run

2.1 Mapping

cd $PATH_mapping.sh
bash mapping.sh

1. publish msg:sensor_msgs/PointCloud2 in topic:/rslidar_points roslaunch rslidar_sdk start.launch
2. run traversability_mapping roslaunch traversability_mapping online.launch
3. move car around the environment
4. save occupancy grid map cd /home/willen/test_ws/src/car_ctr/maps/ rosrun map_server map_saver -f center_map2 map:=/occupancy_map_local

2.2 Run

1. first, move car to the start point of mapping, then log in intranet via Easy Connect
2. run .sh

cd $PATH_run.sh
bash run.sh

3. specify goal point in rviz