Unmanned ground cars (UGCs) can be applied to many areas for assisting humans in various tasks. One critical problem for the control of UGCs is the self-position estimation. The existing GPS-based methods and integration based methods both cannot provide accurate position estimation. Besides, the random loss of sensor signals such as the GPS introduces extra challenges for the problem. This paper proposes a multi-sensor fusion based method to achieve the accurate position estimation for an UGC and also investigates the position estimation with and without the lost position estimation under random loss of GPS signals. The designed methods were implemented on our developed UGC system. The experimental results illustrate the effectiveness of the designed methods and a discussion on the estimation is conducted.
Download Full PDF Version (Non-Commercial Use)