High-Density Automated Valet Parking with Relocation-Free Sequential Operations

Bon Choe; Minhee Kang; Heejin Ahn

eess.SYMar 25, 2026

High-Density Automated Valet Parking with Relocation-Free Sequential Operations

Authors: Bon Choe, Minhee Kang, Heejin Ahn

Organizations: School of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea

Abstract

In this paper, we present DROP, high-Density Relocation-free sequential OPerations in automated valet parking. DROP addresses the challenges in high-density parking & vehicle retrieval without relocations. Each challenge is handled by jointly providing area-efficient layouts and relocation-free parking & exit sequences, considering accessibility with relocation-free sequential operations. To generate such sequences, relocation-free constraints are formulated as explicit logical conditions expressed in boolean variables. Recursive search strategies are employed to derive the logical conditions and enumerate relocation-free sequences under sequential constraints. We demonstrate the effectiveness of our framework through extensive simulations, showing its potential to significantly improve area utilization with relocation-free constraints. We also examine its viability on an application problem with prescribed operational order. The results from all experiments are available at: https://drop-park.github.io.

Figures & tables

Fig. 1 : Comparison of the proposed framework DROP with conventional parking lot and high-density parking.

Fig. 2 : Overview of our DROP framework.

Parameter		Value
Parking lot entrance		Entire left boundary of the lot
Parking stall size $(a\times b)$		3.0 m $\times$ 9.5 m
Vehicle	Size	2.5 m $\times$ 9.0 m
	Front/rear overhang	1.885 m / 2.875 m
	Wheelbase	4.240 m
	Steering angle range	$[-\pi/6,\pi/6]$

TABLE I : Common Experiment Setup Configurations

Fig. 3 : Unique parking lot layouts for the instance 15x12 .

Fig. 4 : Adjacency graph for unique layouts for the instance 15x12 .

Fig. 5 : Accessibility conditions in each layout for the instance 15x12 .

	Layout 1	Layout 2	Layout 3
$\|\textit{exitSeqs}_{\omega}\|$	56	34	1

TABLE II : Relocation-free exit sequences of instance 15x12

Instance	Steps	Brute-Force	DROP
15x12	Layout Gen.	0.011 $\pm$ 0.003
	Postprocess	–	0.0630 $\pm$ 0.0003
	Access. Analysis	–	0.3126 $\pm$ 0.0090
	Seq. Enum.	2.1961 $\pm$ 0.0791	0.0042 $\pm$ 0.0008
	Total	2.2071 $\pm$ 0.0790	0.3908 $\pm$ 0.021
20x16	Layout Gen.	0.026 $\pm$ 0.005

TABLE III : Computation time comparison

Fig. 6 : Example parking (1 st row) and exit (2 nd row) sequences for layout 2 of 15x12 with

\pi=[4,0,1,2,3]

\textit{parkSeq}=[4,2,3,1,0]

and

\textit{exitSeq}=[0,4,2,3,1]

. The current vehicle at each turn is colored in yellow (parking) and green (exiting).

Operation Schedule $\pi$	Layout 1	Layout 2	Layout 3
$[0,1,2,3,4]$	8	2	0
$[1,2,3,4,0]$	24	2	0
$[2,3,4,0,1]$	48	4	0
$[3,4,0,1,2]$	40	12	0
$[4,0,1,2,3]$	16	26	0
Note: Each cell represents $\|\Sigma_{\pi}(\omega,\mathcal{E})\|$ for each layout $\omega$ and operation schedule $\pi$ .

TABLE IV : The results from Solution 2 for instance 15x12

Explore similar work

DriftParking: Trajectory Modeling via Drifting Field for End-to-End Automated Parking
Sep 7, 2026Ziyan Wang, Dong Li, Weibo Wang +4Autonomous Parking Trajectory Generation
Aptiv China (based in Suzhou and Shanghai)
ParkSense: Where Should a Delivery Driver Park? Leveraging Idle AV Compute and Vision-Language Models
Apr 9, 2026Die Hu, Henan Li
University of California, Santa Barbara · Independent Researcher
SunnyParking: Multi-Shot Trajectory Generation and Motion State Awareness for Human-like Parking
Feb 25, 2026Jishu Miao, Han Chen, Jiankun Zhai +4Autonomous Parking Autonomous Driving Planning
Jishu Miao, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi are with Chubu University, Kasugai, Japan. · Han Chen is with Tongji University, Shanghai, China. · Jiankun Zhai, Qi Liu are independent researchers.

Sep 7, 2026J/K move · Enter open · S save

DriftParking: Trajectory Modeling via Drifting Field for End-to-End Automated Parking

Ziyan Wang, Dong Li, Weibo Wang, Yinyin Lu+3

Aptiv China (based in Suzhou and Shanghai)

Autonomous Parking Trajectory Generation End-to-End Autonomous Driving

Open paper AskPDF arXiv

Automated parking requires generating complete and executable trajectories in highly constrained spaces with low tolerance for goal pose error. Existing end-to-end parking methods struggle to jointly achieve inference efficiency, trajectory quality, and precise endpoint alignment, while conventional imitation objectives provide limited supervision on structured deviations from expert maneuver geometry. We propose DriftParking, a one-step trajectory generation framework that reconstructs the drifting-field paradigm for high-precision conditional trajectory generation. Specifically, we replace distribution-level attraction with conditional one-to-one attraction toward the paired expert trajectory, introduce expert-centered constructive repulsion, and adaptively attenuate repulsion near convergence. We further formulate trajectory generation in an endpoint-residual space by decomposing each trajectory into a start-to-goal baseline and a learnable residual, turning endpoint alignment into a representation-level structural constraint on the supervision target while providing a structured space for repulsive supervision. DriftParking achieves state-of-the-art performance across all evaluation metrics. Closed-loop on-vehicle experiments across diverse parking scenarios further show a 97% parking success rate, demonstrating strong zero-shot generalization.

Abstract

Explore similar work

DriftParking: Trajectory Modeling via Drifting Field for End-to-End Automated Parking

ParkSense: Where Should a Delivery Driver Park? Leveraging Idle AV Compute and Vision-Language Models

SunnyParking: Multi-Shot Trajectory Generation and Motion State Awareness for Human-like Parking