Utility-based frontier exloration.

https://www.youtube.com/watch?v=YpDb5tQewAs

https://www.youtube.com/watch?v=MQJ5M0dEo20

https://www.youtube.com/watch?v=X4mQs9AEAl8

• Clone or download to local workspace and build them.
• Make sure the environment is sourced.
• Launch gazebo, then run nav_mapping and demo.py

• tested with ros kinetic
• skimage

roslaunch nav_demos/launch/gazebo.launch

roslaunch nav_demos/launch/nav_mapping.launch

rviz

rosrun my_frontier demo.py

• obtain a map (occupancy grid) and get candidates (frontiers) to be explored
• calculate the best goal to go/explore based on criteria including disance, information gain, and utility
• plan a path from the current pose to the goal
• move the robot and map the environment

• Gmapping is a 2D SLAM implemetation and only 2D map can be created
• Gmapping will be slow when the map grows large