About the project

LFE (Last Frontier Exploration). The original autonomous exploration algorithm computed the distance from the robot to frontiers in a map. However, this approach had a drawback: if the robot had already started planning a path to a frontier and a new, closer frontier appeared during the remapping, the robot would not adapt its path to reach the closer frontier. This led to suboptimal exploration and inefficient use of resources. To address this issue, I made changes to the code in explore.h, explore.cpp, frontier-search.h, and frontier- search.cpp. The key idea was to shift the focus of the cost function from the robot’s current position to the position of the last pursued frontier goal. This allowed the robot to continuously adapt its path to the nearest frontiers, providing a more efficient exploration strategy. So now if during a remmapping a new path needs to be computed it will be computed taking into account the distance from the last frontier and not the distance from the robot, making then that the following frontier chosen will be on the line of the exploration, not leading to jumps from one point to another on the map.

Github repository: