Trajectory Generation
Overview
Trajectory generation studies how to produce or predict sequences of positions and motions that meet a goal while fitting the environment and the moving agent. Methods range from splines, graph search, model-predictive control, and numerical optimization to learned models such as diffusion, flow matching, transformers, and reinforcement learning. Work spans robot navigation and manipulation, autonomous driving, coordinated fleets, and synthetic human mobility, often conditioning plans on maps, language, dynamics, or other agents. A central challenge is balancing speed and diversity with physical feasibility and safety, while testing whether generated trajectories remain realistic and reliable in new settings.
Latest papers
182 papers31 papers in the last four weeks, up 55% on the four weeks before. 0.2% of all new papers.