ADMM-based Continuous Trajectory Optimization in Graphs of Convex Sets
The episode discusses a paper presenting ACTOR, an ADMM-based solver for continuous trajectory optimization in non-convex environments. The authors use polynomial parameterization and a spatio-temporal allocation graph to jointly optimize spatial and temporal decisions. The method offers robustness from naive initializations and faster performance for real-time applications.