Title: Design and Implementation of an Autonomous UGV for the Twenty Second Intelligent Ground Vehicle Competition

Year of Publication: Sep - 2015
Page Numbers: 13-20
Authors: Raghad Al-Harasis , Enas Al-Zmaily, Hamzeh Al-Bishawi, Jamille Abu Shash, Moath Shreim, Belal H. Sababha
Conference Name: The International Conference on Software Engineering, Mobile Computing and Media Informatics (SEMCMI2015)
- Malaysia

Abstract:


The onset of Unmanned Ground Vehicles (UGVs) date back to World War II. These full autonomous robots or remote-controlled robots provide many services for military purposes. The deployment of UGVs in battlefield keeps the soldiers safe from harm, navigates the target points and works as a path tracker. As time goes by, researchers are encouraged to apply UGVs in other domains as in the industrial, road services and the urban domains. The autonomy of UGVs comes from the collective sensory resources and the manipulators that are used to perform specialized tasks. This paper presents the design of a fully autonomous vehicle called “E500”, which has been implemented to compete in the 22nd Intelligent Ground Vehicle Competition (IGVC), held at Oakland University, Rochester, Michigan in June 2014. The E500's body and chassis are custom made. Its power plant is based on two scooter electric motors that are driven through Pulse Width Modulation (PWM). It receives the information from a camera, few ultrasonic range finding sensors and global positioning system (GPS) receiver. The unmanned vehicle also incorporates vision and navigation systems. They are implemented to meet the design requirements of the IGVC competition. The E500's vision system acquires the images through a Microsoft camera then processes them on an onboard laptop. The vehicle was able to extract the features of the road, detecting the white lines and the position of obstacles then figure out the best path to avoid collisions. A navigation algorithm has been developed to achieve high accuracy up to 10 cm using a Samsung mobile phone running android. The algorithms were tested in a green area with two white lines and some obstacles distributed in a random way.