This is a ROS package comprising of the rulah_navigation_goals ROS node software. It was developed at the Spring of 2018 by Georgios Kamaras as a complementary software to his undergraduate thesis in Embodied Artificial Intelligence. More specifically, ugv_navigation_goals consists of an offline path planner for incline terrain navigation. This planner takes as input facts about a specific incline terrain region, like degree of slope, the robot's start position and its goal position and lethal obstacles that exist in this terrain region and outputs a vector of waypoints that have been deduced as the recommended for the robot to traverse the incline terrain region with safety. This software package serves as a proof of concept for its authors proposal regarding the utilization of Bezier curves to construct a smooth curved path from the robot's start position, on the start of the slope, to its goal position, significantly higher in the slope.
- To use this software to its full extend, your robot should run on the ROS middleware.
- Prior to launching the rulah_navigation_goals node you should have already launched your robot's specific packages and especially its move_base node.
Type:
roslaunch rulah_navigation_goals rulah_navigation_goals.launch
For the experimental validation of our work we used the Clearpath Husky UGV in simulation and we created three simulation environments, which can be found on the address https://github.com/yorgosk/husky_simulator.
For the simulations, the Husky Gazebo ROS package was used, with the husky_roboskel_playground_world.launch launch file. The robot was placed in the start position manually, using the teleop_twist_keyboard ROS package, and its move_base was initialized using the Husky Navigation ROS package, with the launch file move_base_mapless_demo.launch. We used the present ugv_navigation_goals package by calling the Rulah Navigation Goals ROS node with the rulah_navigation_goals.launch launch file.
- Georgios Kamaras