Implementation of a 4D fast SLAM including volumetric sum of the UAV

摘要

Traditionally, Simultaneous Localization and Mapping (SLAM) involved a development of map using the actual map coordinates, with the robot itself occupying some 3-dimensional space. SLAM forms an indispensable part of navigation system of an autonomous system. In this paper, we try to assess a method for implementation of a full 4D (4 degree of freedom) solution to SLAM using range and bearing sensors LIDAR (light detection and ranging) and camera for indoor navigation and GPS for outdoor navigation coupled with an Inertial Navigation System (INS) which include inertial sensors such as gyroscopes and accelerometer along with a magnetometer. The method for the Unmanned Aerial Vehicle (UAV) is further modified through volumetric addition into Landmarks and thus into map, in our SLAM procedure in order to decrease the computational complexities and also increase the efficiency of the collision detection algorithms. The UAV motions are restricted to the motion in the three degrees of freedom and yaw, while the pitch and roll are neglected as the UAV is always restricted to move at a constant altitude for a stipulated time-period. This method is significant as it serves for UAV's as Fast SLAM served for ground vehicles.