MultiPush: Learning to Rearrange with Teams of Car-Like Pushers

Jeeho Ahn; Christoforos Mavrogiannis

cs.ROSep 22, 2026

MultiPush: Learning to Rearrange with Teams of Car-Like Pushers

Authors: Jeeho Ahn, Christoforos Mavrogiannis

Organizations: Robotics Department, University of Michigan, Ann Arbor, USA

Abstract

We focus on the problem of rearranging multiple objects within a constrained workspace via pushing using a team of car-like robots. While the use of multiple robots offers the potential for more efficient execution, the need for conflict resolution and the kinematic constraints arising from physics, robot design, and the workspace boundary make this problem especially challenging. Our key insight is that by exploiting the structure introduced by the car-like kinematics of the domain, we could relax the problem into an ordered assignment of Dubins curves to robots. To this end, we introduce MultiPush, a reinforcement-learning based framework that jointly determines an efficient schedule of pushing tasks and their allocation to available robots by leveraging a constraint-aware traversability graph. Across extensive simulated trials with up to 14 objects and teams of two to four robots, MultiPush reduces the makespan by up to 16% compared to the baselines while requiring up to 2.9 times faster planning time. We demonstrate MultiPush on a real-world scenario involving the rearrangement of 12 objects by two and three robots (1/10-scale racecars) in a constrained space.

Figures & tables

Fig. 1 : Overview of the MultiPush framework. The single-robot task sequence produced by ReloPush-BOSS [ 3 ] is expanded into Seed of resting poses, transit times, task durations, and path traces, which parameterizes a pseudo-schedule : a free-space approximation of the multirobot timetable. A learned policy constructs candidate assignment sequences by rolling out against this pseudo-schedule, without invoking the motion planner, and the prioritized multiagent planner verifies them together with the greedy allocation. The best feasible plan is executed. $\eta$ denotes the number of possible actions at each step.

Panel of Figure 2 — Fig. 2 : The push-traversability graph. (a) Each vertex represents a start pose or a goal pose at the objects’ initial or goal poses, respectively. Each edge represents a feasible transfer path (or a series of partial transfer paths) between the start/goal poses, which are composed of Dubins curves. (b) Feasible rearrangements can be found over graph search at each assignment of tasks to robots. Our goal is to find a feasible assignment sequence with a low makespan.

Component	Description
States from the seed (one set per instance)
Resting poses	robot initial poses and the pose where each task’s final transfer path ends
Transit times	free-space Reeds–Shepp [ 31 ] time from each resting pose to each task’s pushing pose
Task durations	each task’s busy time in the seed
Path traces	sampled waypoints of each task’s paths; used for geometry only
Precedence relations	task pairs kept in seed order because a start or goal placement blocks a transfer

TABLE I: The geometry, durations, and precedence relations fixed by the seed, and the pseudo-schedule

\hat{\Gamma}_{h}

advanced over them (Sec. IV-B ). The upper block is computed once per instance from the seed and the robots’ initial poses, and the lower block is updated at each assignment. Transit times ignore the other robots, which is the surrogate’s central approximation.

Feature	Description
$\max(0,\ \hat{C}_{a}-\hat{C}_{\text{max}})$	extension of the pseudo-schedule
$\max(0,\ \hat{C}_{\text{max}}-\hat{C}_{a})$	margin ahead of $\hat{C}_{\text{max}}$
$\mathrm{transit}$	free-space time to the pushing pose
$\hat{C}_{i}-\min_{i^{\prime}}\hat{C}_{i^{\prime}}$	delay behind the earliest free robot
$\max_{i}\hat{C}^{+}_{i}-\min_{i}\hat{C}^{+}_{i}$	spread of availabilities
$\sum_{j}\mathrm{blockage}_{j}$	objects on the corridor

TABLE II: Features of

\phi(s_{h},a_{h})

, evaluated for every candidate assignment from the seed and

\hat{\Gamma}_{h}

, without path planning. The availability

\hat{C}_{i}

is the time robot

i

completes its last assigned task in

\hat{\Gamma}_{h}

, and

\hat{C}_{\text{max}}=\max_{i}\hat{C}_{i}

before the assignment. The prediction

\hat{C}_{a}

adds the candidate’s transit time and task duration from the seed to its robot’s availability, and

\hat{C}^{+}_{i}

are the availabilities after the assignment. Clearance is the smallest boundary distance over the candidate’s pushing poses, normalized by the turning radius

r

(one at the boundary, zero at

r

or more). The last feature counts earlier-ordered tasks in

\sigma_{0}

whose undelivered objects obstruct the non-push segments of the candidate task in the seed.

Fig. 3 : Qualitative results. The first column shows four scenarios drawn from the test set involving respectively

m\in\{8\}

, 10, 12, and 14 objects; objects’ start poses are drawn in yellow, and goal poses in white. The remaining columns show the resulting trajectories planned by MultiPush for

n=1

2

3

, and

4

robots, colored by robot.

		$4.5\times 5.5$ m 2 (trained workspace)				$5.0\times 5.0$ m 2 (unseen workspace)
		Rel. Makespan				Rel. Makespan
$n$	Method	vs. GREEDY	vs. BOSS	Planning (s)	Success (%)	vs. GREEDY	vs. BOSS	Planning (s)	Success (%)
1	BOSS	–	1.000	–	100.0	–	1.000	–	100.0
1	MultiPush+BOSS	–	0.947 (0.044)	5.1 (4.7)	100.0	–	0.945 (0.047)	5.0 (5.2)	100.0
2	GREEDY	1.000	0.602 (0.064)	–	96.0	1.000	0.593 (0.056)	–	93.0
	RANDOM	1.024 (0.133)	0.613 (0.078)	11.9 (8.2)	92.5	1.026 (0.130)	0.606 (0.076)	11.5 (9.5)	92.5

TABLE III : Relative makespan, planning time, and success rate with

n

robots on the held-out test set for the trained and an unseen workspace. Each workspace column covers 400 test instances, 100 per object count

m\in\{8,10,12,14\}

. The test instances appear in neither the training nor the validation set, and ReloPush-BOSS solves every instance. The relative makespan is measured against GREEDY and against the single-robot makespan of BOSS, both averaged over the instances solved by every method. At

n=1

, MultiPush+BOSS verifies the BOSS solution in addition to the

K^{\prime}

candidates in place of the GREEDY allocation.

Fig. 4 : Mean makespan, relative to that of GREEDY, on the validation set across training, warm-started with behavior cloning. Each point averages the relative makespan of the best feasible of the $K^{\prime}$ sampled candidates over the validation instances the policy solves. The star marks the policy selected for inference. The shaded band is the standard error of the mean. The standard deviation across the solved instances is $0.10$ to $0.12$ at every epoch.

Planner evaluations that returned infeasible (%)
	Trained workspace			Unseen workspace
Method	$n=2$	$n=3$	$n=4$	$n=2$	$n=3$	$n=4$	All
RANDOM	40.6	37.7	36.3	37.7	35.3	33.6	36.9
LNS	38.7	42.0	43.4	36.4	40.6	43.2	40.7
MultiPush	31.2	28.3	26.5	29.1	26.7	24.6	27.7

TABLE IV : Fraction of planner evaluations that returned infeasible on the test set of Table III .

Explore similar work

Non-Prehensile Multi-Object Placement with Mobile Robots via Diffusion-Based Trajectory Planning
Oct 8, 2026Lance Pharand, Megnath Ramesh, Mark Chiu +2Non-Prehensile Manipulation Mobile Robotics
Electrical and Computer Engineering Department at the University of Waterloo, ON, Canada
Embodied Scene Rearrangement Planning
Aug 27, 2026Canzhi Chen, Zan Wang, Siqi Zhu +3Robot Task Planning Long-Horizon Robotic Manipulation
School of Computer Science and Technology, Beijing Institute of Technology, Beijing 100081, China
FORMICA: Decision-Focused Learning for Communication-Free Multi-Robot Task Allocation
Feb 20, 2026Antonio Lopez, Jack Muirhead, Carlo PinciroliMulti-Robot Task Allocation Multi-Agent Systems
Dept. of Robotics Engineering, Worcester Polytechnic Institute, Worcester MA 01609, USA

Oct 8, 2026J/K move · Enter open · S save

Non-Prehensile Multi-Object Placement with Mobile Robots via Diffusion-Based Trajectory Planning

Lance Pharand, Megnath Ramesh, Mark Chiu, Yash Vardhan Pant+1

Electrical and Computer Engineering Department at the University of Waterloo, ON, Canada

Non-Prehensile Manipulation Mobile Robotics Diffusion-Based Planning

Open paper AskPDF arXiv

Non-prehensile manipulation with mobile robots has primarily targeted placement in coarse goal regions. Precision placement has been studied with manipulators on benchmarks such as Push-T, where an omnidirectional end effector can readily re-approach an object from any side. For a mobile robot, correcting the same placement errors may require wide repositioning maneuvers around the object. We formalize multi-object pushing with a mobile robot as a multi-target assignment problem at tight tolerances. We propose a diffusion planner with a causal factorization: each robot denoising step is followed by an object step conditioned on the current clean robot estimate. This factorization allows us to sample multiple object outcomes for a single robot plan, capturing the multi-modality of pushing. We also design a set of pushing-specific costs for mobile robots, used for both guidance and best-of-

N

sampling at inference. On Push-T, our planner attains the highest success rate of the methods we compare, while having the lowest rate of implausible object motion. Across three multi-box pushing environments resembling warehouse scenarios, it raises the fraction of boxes placed by

21.5

percentage points over an unguided baseline, and replans

1.8\times

faster than a planner that denoises the two trajectories together. Lastly, we deploy the guided planner on a TurtleBot 3 Burger with object shapes absent from the training data, where it places

85.0\%

of objects within tolerance against Diffusion Policy's

37.5\%

Abstract

Explore similar work

Non-Prehensile Multi-Object Placement with Mobile Robots via Diffusion-Based Trajectory Planning

Embodied Scene Rearrangement Planning

FORMICA: Decision-Focused Learning for Communication-Free Multi-Robot Task Allocation