Kinodynamic optimal trajectories generation in cluttered environments using Bernstein polynomials
Bruno Strano, Michele Furci, Lorenzo Marconi · Mechatronics · 2023
We develop a trajectory planning strategy for autonomous agents using Bernstein polynomials. The focus is on generating trajectories that are functionally controllable for under-actuated vehicles and that guarantee a safe path to avoid obstacles. The algorithm is suitable for local replanning and onboard quick re-computation of the trajectory when unplanned obstacles pop up. By leveraging on the properties of Bernstein polynomials the problem is cast into a quadratic programming problem that can be solved in polynomial time.