I worked on developing the complete code stack for the autonomous traversal of a prototype Mars Rover for the University, Anatolian and International Rover Challenge.
A* path planning was implemented with the help of pointcloud generated with the help of a ZED stereocamera. The rover was localized in the generated map using Pixhawk flight controller and a GPS module.
I also worked on implementing deep learning techniques including YOLO wich was integrated with darknet_ros using a Jetson Xavier. The YOLO was implemented to detect arrow signs for the rover and follow the directions to reach an end goal. A custum dataset was made to complete this task.
In the URC competition, held in Utah, USA our team concluded the competition with a 2nd rank in Asia and in the ARC competition (Istanbul), we stood second worldwide.