Trajectory Refinement
Trajectory refinement focuses on improving the accuracy and efficiency of predicted or generated robot trajectories, addressing challenges like generalization, robustness to noise, and computational cost. Current research emphasizes methods leveraging diffusion models, transformers, and neural ordinary differential equations to refine trajectories based on various sources of information, including demonstrations, sensor data, and high-level commands. These advancements are crucial for improving the performance of autonomous systems in diverse applications, such as robotics, autonomous driving, and traffic simulation, by enabling more accurate and reliable motion planning and control.
Papers
November 6, 2024
October 30, 2024
October 28, 2024
October 16, 2024
October 10, 2024
October 5, 2024
September 13, 2024
September 9, 2024
June 6, 2024
May 24, 2024
April 29, 2024
March 18, 2024
February 2, 2024
January 29, 2024
January 8, 2024
November 4, 2023
November 2, 2023
September 18, 2023
July 25, 2023