An Optimization-based Motion Planner for Safe Autonomous Driving

Francisco Eiras; Majd Hawasly; Stefano V. Albrecht; Subramanian Ramamoorthy

An Optimization-based Motion Planner for Safe Autonomous Driving

Francisco Eiras, Majd Hawasly, Stefano V. Albrecht, Subramanian Ramamoorthy

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Guaranteeing safety in motion planning is a crucial bottleneck on the path towards wider adoption of autonomous driving technology. A promising direction is to pose safety requirements as planning constraints in nonlinear optimization problems of motion synthesis. However, many implementations of this approach are hindered by uncertain convergence and local optimality of the solutions, affecting the planner’s overall robustness. In this paper, we propose a novel two-stage optimization framework: we first find the solution to a Mixed-Integer Linear Programming (MILP) approximation of the motion synthesis problem, which in turn initializes a second Nonlinear Programming (NLP) formulation. We show that initializing the NLP stage with the MILP solution leads to better convergence, lower costs, and outperforms a state-of-the-art Nonlinear Model Predictive Control baseline in both progress and comfort metrics.

Original language	English
Title of host publication	Second (virtual) workshop on Robust autonomy: Safe robot learning and control in uncertain real-world environments
Publication status	Published - 2020
Externally published	Yes

Access to Document

https://scholar.google.com/citations?view_op=view_citation&hl=en&user=dvTSHIkAAAAJ&pagesize=100&citation_for_view=dvTSHIkAAAAJ:738O_yMBCRsC

Cite this

Eiras, F., Hawasly, M., Albrecht, S. V., & Ramamoorthy, S. (2020). An Optimization-based Motion Planner for Safe Autonomous Driving. In Second (virtual) workshop on Robust autonomy: Safe robot learning and control in uncertain real-world environments https://scholar.google.com/citations?view_op=view_citation&hl=en&user=dvTSHIkAAAAJ&pagesize=100&citation_for_view=dvTSHIkAAAAJ:738O_yMBCRsC

@inproceedings{d62aa71553834fd6b8e5e5c28e1faa7b,

title = "An Optimization-based Motion Planner for Safe Autonomous Driving",

abstract = "Guaranteeing safety in motion planning is a crucial bottleneck on the path towards wider adoption of autonomous driving technology. A promising direction is to pose safety requirements as planning constraints in nonlinear optimization problems of motion synthesis. However, many implementations of this approach are hindered by uncertain convergence and local optimality of the solutions, affecting the planner{\textquoteright}s overall robustness. In this paper, we propose a novel two-stage optimization framework: we first find the solution to a Mixed-Integer Linear Programming (MILP) approximation of the motion synthesis problem, which in turn initializes a second Nonlinear Programming (NLP) formulation. We show that initializing the NLP stage with the MILP solution leads to better convergence, lower costs, and outperforms a state-of-the-art Nonlinear Model Predictive Control baseline in both progress and comfort metrics.",

author = "Francisco Eiras and Majd Hawasly and Albrecht, {Stefano V.} and Subramanian Ramamoorthy",

year = "2020",

language = "English",

booktitle = "Second (virtual) workshop on Robust autonomy: Safe robot learning and control in uncertain real-world environments",

}

Eiras, F, Hawasly, M, Albrecht, SV & Ramamoorthy, S 2020, An Optimization-based Motion Planner for Safe Autonomous Driving. in Second (virtual) workshop on Robust autonomy: Safe robot learning and control in uncertain real-world environments. <https://scholar.google.com/citations?view_op=view_citation&hl=en&user=dvTSHIkAAAAJ&pagesize=100&citation_for_view=dvTSHIkAAAAJ:738O_yMBCRsC>

TY - GEN

T1 - An Optimization-based Motion Planner for Safe Autonomous Driving

AU - Eiras, Francisco

AU - Hawasly, Majd

AU - Albrecht, Stefano V.

AU - Ramamoorthy, Subramanian

PY - 2020

Y1 - 2020

N2 - Guaranteeing safety in motion planning is a crucial bottleneck on the path towards wider adoption of autonomous driving technology. A promising direction is to pose safety requirements as planning constraints in nonlinear optimization problems of motion synthesis. However, many implementations of this approach are hindered by uncertain convergence and local optimality of the solutions, affecting the planner’s overall robustness. In this paper, we propose a novel two-stage optimization framework: we first find the solution to a Mixed-Integer Linear Programming (MILP) approximation of the motion synthesis problem, which in turn initializes a second Nonlinear Programming (NLP) formulation. We show that initializing the NLP stage with the MILP solution leads to better convergence, lower costs, and outperforms a state-of-the-art Nonlinear Model Predictive Control baseline in both progress and comfort metrics.

AB - Guaranteeing safety in motion planning is a crucial bottleneck on the path towards wider adoption of autonomous driving technology. A promising direction is to pose safety requirements as planning constraints in nonlinear optimization problems of motion synthesis. However, many implementations of this approach are hindered by uncertain convergence and local optimality of the solutions, affecting the planner’s overall robustness. In this paper, we propose a novel two-stage optimization framework: we first find the solution to a Mixed-Integer Linear Programming (MILP) approximation of the motion synthesis problem, which in turn initializes a second Nonlinear Programming (NLP) formulation. We show that initializing the NLP stage with the MILP solution leads to better convergence, lower costs, and outperforms a state-of-the-art Nonlinear Model Predictive Control baseline in both progress and comfort metrics.

M3 - Conference contribution

BT - Second (virtual) workshop on Robust autonomy: Safe robot learning and control in uncertain real-world environments

ER -

An Optimization-based Motion Planner for Safe Autonomous Driving

Abstract

Access to Document

Fingerprint

Cite this