10/03/2026
Proudly presenting our qualifying run for the Bosch Future Mobility Challenge 2026.
In this run, our autonomous vehicle successfully navigates a wide range of real competition scenarios including stop signs, crosswalks, pedestrians, intersections, roundabouts, highways, and overtaking all on the BFMC track.
Our system is built around a topological localization approach, where traffic sign locations are integrated directly into a graph-based map of the track. Using this structure, the system generates the route, curvature, and the nearest traffic sign information along the path. The vehicle primarily follows the road using lane masking and centerline-based pure pursuit control, while the sign-based topological localization continuously tracks the carโs position on the track and ensures correct decision-making at key points.
Our technology stack combines the power of ROS 2 Humble, the BFMC Brain coordination system with WebUI, and a STM32 Nucleo microcontroller responsible for low-level vehicle control. The perception, planning, and decision-making modules run as ROS2 nodes, communicating through DDS (Data Distribution Service), while the microcontroller handles motors, steering, and sensor interfaces.
Despite working with limited hardware, we focused on building a robust and practical system. Our vehicle operates without UWB localization or high-end positioning systems, relying instead on computer vision, traffic sign understanding, and our topological localization strategy. This makes the system lightweight, efficient, and accessible for teams with limited resources.
This project represents countless hours of sleepless nights, coding, testing, debugging, and tuning by the dedicated members of Autodrive. Every line of code and every iteration brought us closer to a reliable autonomous system.
We are proud to share this milestone and hopeful that, In Sha Allah, our work will earn us a place in the next stage of the competition.
๐ Team Autodrive
