Time-bounded lattice for efficient planning in dynamic environments

摘要: For vehicles navigating initially unknown cluttered environments, current state-of-the-art planning algorithms are able to plan and re-plan dynamically-feasible paths efficiently robustly. It is still a challenge, however, deal well with the surroundings that both highly dynamic. Planning under these conditions more difficult for two reasons. First, tracking predicting trajectories of moving objects (i.e., cars, humans) very noisy. Second, process computationally expensive because increased dimensionality state-space, time as an additional variable. Moreover, re-planning needs be invoked often since obstacles need constantly re-estimated. In this paper, we develop path algorithm addresses challenges. choose representation dynamic models their predicted uncertainty associated predictions. provide real-time guarantees on performance obstacles, propose utilize novel data structure - time-bounded lattice merges together short-term in longterm without time. We demonstrate effectiveness approach simulations up 30 real robots.