The "Automated and Connected Driving Challenges (ACDC) – Research Project" at RWTH Aachen University's Institute of Automotive Engineering (IKA) focuses on contemporary research in automated and connected driving. This course allows students to engage in a research project, contributing to ongoing challenges in the field. It emphasizes the application of theoretical knowledge, fosters independent group work, and encourages active participation in shaping the future of Intelligent Transport Systems (ITS) that are safe, sustainable, and accessible.
For more information, visit the ACDC Research Project page.
- Accessibility of Research: Utilizing scaled-down robotic models of automated vehicles makes teaching, research, and development more accessible.
- Transferability of Knowledge: The algorithms and insights gained from these models can often be applied to full-size vehicles.
- Foundation for Advanced Development: Implementing basic self-driving functionalities paves the way for more advanced research and functional development tasks.
The primary task was to develop a methodology for visually detecting, tracking, and following driving lanes using a scaled automated and connected vehicle. This project involved several key objectives:
- Lane Detection: Creating algorithms capable of identifying lane markings in various environmental conditions.
- Lane Tracking: Implementing methods for continuously tracking the detected lanes to ensure stable vehicle guidance.
- Lane Following: Integrating the detection and tracking capabilities to enable the vehicle to autonomously follow the lane in real-time.
- Develop a robust visual processing pipeline for lane detection using camera data.
- Test and refine algorithms to ensure reliability across diverse scenarios, including different lighting and weather conditions.
- Implement feedback mechanisms for lane tracking that enhance the vehicle's autonomous navigation capabilities.
The research project employed a systematic approach:
- Data Collection: Capturing video data from various driving scenarios to train and test the lane detection algorithms.
- Algorithm Development: Utilizing computer vision techniques and/or machine learning to process the visual data and accurately detect lane markings.