Odometry of a crawler robotic platform

This project presents the main characteristics of a two sets of hardware-software architecture and the implementation of visual odometry and mapping for controlling the autonomous mobile robot "Killer Crawler" in real environments. This vehicle is a rover-type research platform with dual-wheel motorize, which operates in rough terrain under arbitrary conditions. The whole system is designed to allow easy reconfiguration and adaption for new tasks. The hardware architecture, has a dedicated computer based in OS Ubuntu 12.10, where all the functions of the vehicle are controlled using cameras for data acquisition and hardware drivers to link sensors and actuators. The objective of this thesis is develop a visual odometry and mapping method for an autonomous platform called Killer Crawler, that includes the design of hardware and software based in low level programming, using a combination between different sensors and a cameras system.

Hardware Configuration:

• Embedded CPU Kontron ComExpress AI (Intel core i7, XV1 Board by EKF systems)

• Stereo cameras Allied Synergy F-125B/C (1292 x 964 pixels)

• Ardupilot APM 2.0

• Traco Power voltage regulator (3.3v, 5.0v and 12v)

• Pant-tilt system

• Battery level indicator

Software Configuration:

• Mission Planner 10

• Ground-station for remote video monitoring

• Real-time D6DSLAM [Patent]