Mapping, localization, and path-planning are three fundamental problems of robotic.
Robot needs a map to perform actions like path-planning. When positioning system is
not available, the map is also used for localization. A lot of researches have been done
in this area. And newly emerging ranging sensors, like Kinect and TOF camera, have
widen people’s choices and greatly enhanced innovative ideas in robot mapping. The
price of these sensors is not very high and the performance is decent, which makes low
cost, high performance mobile robot solution possible.
In this thesis, different existing state of the art mapping methods are studied. Based on
literature studies, different ranging sensors for mapping are evaluated. And by using the
3D ranging sensor, three mapping methods are implemented. Occupancy grid map with
scan matching, feature-grid hybrid map with map pruning and simple points map with
ICP algorithm. Basic potential field path-planning is also implemented.
The experiment results illustrate the performance of each method.