Abstract
Applying unmanned ground vehicle (UGV) to complex environments, it is important to control not only the mechatronic system but also the precision of localization and trajectory tracking of the UGV. This study introduces simultaneous localization and mapping (SLAM). Based on RTAB-Map, the shortcomings of vision odometry in RGB-D camera has been overcome by adding inertial measurement unit and GPS-RTK sensor fusion, reducing the static positioning deviation from ±5.2 mm to ±3 mm under interference. Regarding trajectory tracking, the model predictive control can enable the UGV precisely deal with the rugged and changeable environment with the above mentioned high-precision SLAM positioning which is mentioned above.