Exploration and Path Planning

The black marked cells are obstacles, the light grey cells are free space and the dark grey cells are unknown terrain. If there are obstacles in the map the robot is able to estimate < it's position and it's moving velocity relatively to these obstacles. Due to the precision of the laser scanner there is no odometry from the robot required. If the robot is not remote controlled a second problem has to be solved. The robot has autonomously to choose where to go next. To solve this problem the hector_exploration_planner is used. The planner generates goals and paths to these goals. The hector_exploration_planner package is a library that can be used inside as ROS node.

Fig. 1: Exploration of one of
the University laboratories

A path will be generated by calculating the cost to reach a closed frontier. A frontier is a cell in the occupancy grid map that is marked as free space and has a neighboring cell that is marked as unknown terrain. Costs are the distance to the frontier and the safety of reaching the frontier.With a weighting factor it can be adjusted how much a safer path is preferred over a shorter one.

The map submitted after each mission will generated with the hector_geotiff ROS package. It provides a node that can be used to save occupancy grid map, robot trajectory and object of interest data to RoboCupRescue compliant GeoTiff images.