Hello, I'm Rwik Rana.
A Robotics Researcher
I am a Ph.D. student at the University of Texas at Austin, working with Prof. Joydeep Biswas in the Autonomous Mobile Robotics Laboratory (AMRL).
About Me
My research focuses on the intersection of model-based control and model-free reinforcement learning. I develop algorithms to improve the performance and reliability of autonomous systems like drones and off-road vehicles, with a focus on trajectory tracking, control optimization, and perception.
Publications & Patents
Projects
Deep Car Acrobatics
Developing a forward prediction model for sampling-based planners to better handle complex vehicle dynamics, including air-borne maneuvers and collisions.
Learn More →Off-Road Navigation (UT ARL)
Contributed to the OVerSeeC system, which generates costmaps from satellite imagery and natural language to assist global planning for off-road vehicles.
Learn More →Fast Off-Road Navigation (DARPA RACER)
Worked on an uncertainty-aware semantic segmentation model (BEVNet) for high-speed off-road driving, which showed an 18% improvement in local planner performance.
Learn More →Tactile Sensor Super Resolution
Designed a VAE-based spatio-temporal architecture to increase tactile sensor resolution by a factor of four.
Learn More →Memory Guided Road Detection
Developed a road segmentation framework that balances accuracy and speed by interleaving feature extractors, resulting in a 3x inference speed-up.
View Repository →End-to-end Autonomous Racing
Designed an end-to-end racing algorithm for the CARLA simulator, using a VAE with a genetic algorithm for trajectory optimization.
View on ArXiv →Get In Touch
I'm always open to discussing new projects, creative ideas, or opportunities. Feel free to reach out.