Intelligent Vehicle Trajectory Planning via Combined Dynamic and Quadratic Programming

To address the challenges posed by non-convex optimization in path planning for intelligent vehicles using numerical optimization techniques, this study introduces a combined approach leveraging dynamic programming (DP) and quadratic programming (QP). The dynamic programming algorithm is employed for path search, with the solution space defined by road and obstacle boundaries. Subsequently, a QP algorithm is applied to optimize the path represented by a quintic polynomial curve, effectively mitigating the no-solution issue commonly associated with non-convex optimization. Similarly, velocity planning follows a dual approach of dynamic programming and quadratic programming. The dynamic programming algorithm generates an initial velocity solution, incorporating vehicle dynamics constraints, and the QP algorithm is then used to refine the velocity profile. Simulations demonstrate the efficacy of the proposed trajectory planning strategy in both static and dynamic scenarios.