Tunable Trajectory Planner Using G³ Curves

Trajectory planning is commonly used as part of a local planner in autonomous driving. This paper considers the problem continuous-curvature-rate trajectory between fixed start and goal states that minimizes tunable trade-off passenger comfort travel time. The an instance infinite dimensional optimization over two continuous functions: path velocity profile. We propose simplification this facilitates discretization both functions. also proposes method to quickly generate minimal-length paths based on single tuning parameter: second derivative curvature. Further, we discretize set profiles along given into selection longitudinal jerk way-points path. Finally, repeatedly solve iterative fashion. Numerical examples are provided illustrate benefits proposed methods.