Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

OCEAN: An Openspace Collision-free Trajectory Planner for Autonomous Parking Based on ADMM

Dongxu Wang, Yanbin Lu, Weilong Liu, Hao Zuo, Jiade Xin, Xiang Long, Yuncheng Jiang

TL;DR

This work tackles autonomous openspace parking under non-convex collision constraints by formulating a Model Predictive Control problem with full-dimension vehicle dynamics and a dual convex reformulation of collision avoidance. OCEAN leverages ADMM to decompose the problem into parallelizable subproblems of type QP, SQP, and SOCP, solving them efficiently without heavy reliance on warm starts. The approach is validated through hundreds of simulations and real parking tests, showing robust performance and substantial reductions in planning time compared with benchmark methods. The results indicate that real-time, scalable autonomous valet parking is feasible on platforms with modest computing power, enabling practical deployment.

Abstract

In this paper, we propose an Openspace Collision-freE trAjectory plaNner (OCEAN) for autonomous parking. OCEAN is an optimization-based trajectory planner accelerated by Alternating Direction Method of Multiplier (ADMM) with enhanced computational efficiency and robustness, and is suitable for all scenes with few dynamic obstacles. Starting from a hierarchical optimization-based collision avoidance framework, the trajectory planning problem is first warm-started by a collision-free Hybrid A* trajectory, then the collision avoidance trajectory planning problem is reformulated as a smooth and convex dual form, and solved by ADMM in parallel. The optimization variables are carefully split into several groups so that ADMM sub-problems are formulated as Quadratic Programming (QP), Sequential Quadratic Programming (SQP),and Second Order Cone Programming (SOCP) problems that can be efficiently and robustly solved. We validate our method both in hundreds of simulation scenarios and hundreds of hours of public parking areas. The results show that the proposed method has better system performance compared with other benchmarks.

OCEAN: An Openspace Collision-free Trajectory Planner for Autonomous Parking Based on ADMM

TL;DR

This work tackles autonomous openspace parking under non-convex collision constraints by formulating a Model Predictive Control problem with full-dimension vehicle dynamics and a dual convex reformulation of collision avoidance. OCEAN leverages ADMM to decompose the problem into parallelizable subproblems of type QP, SQP, and SOCP, solving them efficiently without heavy reliance on warm starts. The approach is validated through hundreds of simulations and real parking tests, showing robust performance and substantial reductions in planning time compared with benchmark methods. The results indicate that real-time, scalable autonomous valet parking is feasible on platforms with modest computing power, enabling practical deployment.

Abstract

In this paper, we propose an Openspace Collision-freE trAjectory plaNner (OCEAN) for autonomous parking. OCEAN is an optimization-based trajectory planner accelerated by Alternating Direction Method of Multiplier (ADMM) with enhanced computational efficiency and robustness, and is suitable for all scenes with few dynamic obstacles. Starting from a hierarchical optimization-based collision avoidance framework, the trajectory planning problem is first warm-started by a collision-free Hybrid A* trajectory, then the collision avoidance trajectory planning problem is reformulated as a smooth and convex dual form, and solved by ADMM in parallel. The optimization variables are carefully split into several groups so that ADMM sub-problems are formulated as Quadratic Programming (QP), Sequential Quadratic Programming (SQP),and Second Order Cone Programming (SOCP) problems that can be efficiently and robustly solved. We validate our method both in hundreds of simulation scenarios and hundreds of hours of public parking areas. The results show that the proposed method has better system performance compared with other benchmarks.
Paper Structure (20 sections, 33 equations, 5 figures, 5 tables, 1 algorithm)

This paper contains 20 sections, 33 equations, 5 figures, 5 tables, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORK
  3. Ellipsoids or Sphere Approximation
  4. Safe Corridor
  5. Double Signed Distance Obstacle Representation
  6. Linear Programming
  7. Euclidean Signed Distance Field
  8. Accelerate Optimization Algorithms
  9. PROBLEM FORMULATION
  10. Vehicle Dynamic Model
  11. Full Dimension Modeling and Collision Avoidance
  12. Problem Statement
  13. COLLISION-FREE TRAJECTORY PLANNER
  14. Parallel Computation
  15. SIMULATION AND RESULTS
  16. ...and 5 more sections

Figures (5)

  • Figure 1: Benchmark Scenarios.
  • Figure 2: Parking Trajectory Visualization and Kinematic Parameters Sampling.
  • Figure 3: Planning trajectories in simulation including scenarios: 1) oblique; 2) vertical; 3) parallel.
  • Figure 4: Comparison of Time Consumption in Different Solution Steps for Three Algorithms.
  • Figure 5: Real Test Tracking Trajectory Visualization.