Dynamic Trajectory Planning for Autonomous Driving Based on Fluid Simulation

This paper describes the use of Lattice Boltzmann Method for dynamic autonomous driving trajectory planning. By simulating fluid flow on roads as a two-dimensional tubes in a small area around ego car, a vector map is created which can be used as a direct basis for driving trajectory. To tailor the phenomena of fluid flow for legal driving trajectory, the simulated fluid source is placed at an offset angle while dynamically following target car and road borders generate additional potential field. Experiments have shown that fluid simulation needs to propagate relatively to car driven distance for the algorithm to be feasible. During evaluation such ratio has been found and approximated.