I was lucky enough to participate in the 2024 Mars Rover challenge with many talented students. The University Rover Challenge is an annual competition involving over 100 teams from 15 countries. BYU has done well in this competition in the past. In 2024, we scored 3rd place.
Technical details of the BYU 2024 Mars Rover
The University Rover Challenge includes four parts: extreme delivery, autonomous navigation, equipement servicing, and a science task. You can find information about each of the missions here. My part was in the autonomy mission where the rover had to navigate autonomously to various GPS points. The terrain began smooth but quickly entered challenging rocky & hilly terrain. Not only did the rover need to navigate over this bumpy terrain with its mechanical rocker-bogey system but it also needed to avoid larger obstacles and steep hills that it couldn't handle directly. This was a major challenge for us. In addition to navigating this difficult terrain, the rover had to identify and approach different objects such as a water bottle and an orange mallet.
The rover detecting and avoiding obstacles using an obstacle detection and path finding algorithm
In order to address these concerns, we opted for an obstacle detection system paired with an DNN object recognition system. The project itself is built in Python ROS (Robot Operating System) running on an NVidia Jetson Orin. I headed the obstacle detection system. Because this is a project that is passed from one graduating class to another, I started by fixing the old obstacle detection system that hadn't been touched for a few years. It was in pretty rough shape, throwing many false positives at random. The solution came about as I endeavored to understand, organize, and minimize the old code. Despite this being a new subject to me, I was able to understand how the object detection algorithm worked (Used OpenCV) and adjust the parameters to be capable of detecting hazardous objects in our path. Gratefully, the obstacle detection algorithm performed admirably.
I learned a lot during the course of this project. To be frank, I was a little surprised to be accepted into this sought after team. I learned so much from both my experiences as well as the people on the team itself. The dedication, interest, motivation, knowldge, and diligence that they showed have left a lasting impact on me. The project and techincal skills required taught me a lot as well. I had to learn about ROS, local networks, RF transmitters and receivers, (even more) linux commands, vim, OpenCV, blob detection, guassian blurs, sterioscopic cameras, obstacle detection, and working cohesively in a large group. I'm grateful to have learned as much as I did and am proud of my critical work.