Research Project Title:
Vision-Based Approach to Localization and Mapping in Autonomous Vehicles
abstract:Most of today’s fully autonomous research vehicles use a combination of low-cost vision sensors and high-cost LiDAR sensors. Implementations of systems using purely vision have demonstrated that some amount of autonomy is possible using only RGB and depth sensing technology. However, there are several limitations to the vision-based approach. These include the cost of computation and lack of robustness to varying light conditions. My research project at MIT’s Aerospace Controls Laboratory will focus on developing RGB-D vision-based algorithms for localization and mapping. The algorithms developed through this project will address the issues in current solutions, and lead to a scalable, reliable solution for diverse robotic platforms.
Through this SuperUROP, I want to gain more experience in robotics research and make a positive contribution to my lab. Robotics is a very fascinating field to be in. Especially considering the potential it has to improve our lives, I'm excited to bring together ideas from the computer vision and graphics communities and learn more about the problems under this research domain.