Cooperative mapping and self-localization for multiple quadrocopters

Cooperative Simultaneous Localization and Mapping (SLAM) for Multiple Quadrocopters is a method to solve the problem of navigation in a larger area within a short span of time. This research uses tum_ardrone package, a package compatible with Robot Operating System (development platform to build intelligence) and having a capability to provide Monocular SLAM based on the concept of Parallel Tracking and Mapping. For the implementation, two quadrocopters are using this tum_ardrone package and perform local SLAM operation. The data provided by this package is converted into 3D point cloud data for each quadrocopter and the two have their own local maps in the form of 3D point clouds. These local maps are merged to form a global map using point cloud registration technique.;The two quadrocopters can localize themselves in this global map by using the technique of feature matching. During navigation, the quadrocopter captures the image and the features are extracted from the image and converted to point cloud data. This point cloud feature data is compared with the point cloud data in the global map and computes the location in the global map which in turn acts as a sensor input to the estimation filter and gives the location of the quadrocopters in the global map. The global 3D map is successfully obtained in this research work with a decent accuracy and the localization technique gives the location of the quadrocopter based on the global map using which it can navigate in that area.