Can a Robot Walk the Robotic Dog: Triple-Zero Collaborative Navigation for Heterogeneous Multi-Agent Systems

Yaxuan Wang; Yifan Xiang; Ke Li; Xun Zhang; BoWen Ye; Zhuochen Fan; Fei Wei; Tong Yang

cs.ROMar 23, 2026

Can a Robot Walk the Robotic Dog: Triple-Zero Collaborative Navigation for Heterogeneous Multi-Agent Systems

Authors: Yaxuan Wang, Yifan Xiang, Ke Li, Xun Zhang, BoWen Ye, Zhuochen Fan, Fei Wei, Tong Yang

Organizations: Yuanpei College, Peking University · School of Computer Science, Peking University · School of Computer Science, Beijing University of Posts and Telecommunications · Pengcheng Laboratory · Beijing Jinruyi Large Model Technology Co., Ltd.

Abstract

We present Triple Zero Path Planning (TZPP), a collaborative framework for heterogeneous multi-robot systems that requires zero training, zero prior knowledge, and zero simulation. TZPP employs a coordinator--explorer architecture: a humanoid robot handles task coordination, while a quadruped robot explores and identifies feasible paths using guidance from a multimodal large language model. We implement TZPP on Unitree G1 and Go2 robots and evaluate it across diverse indoor and outdoor environments, including obstacle-rich and landmark-sparse settings. Experiments show that TZPP achieves robust, human-comparable efficiency and strong adaptability to unseen scenarios. By eliminating reliance on training and simulation, TZPP offers a practical path toward real-world deployment of heterogeneous robot cooperation. Our code and video are provided at: https://github.com/triple-zeropp/Triple-zero-robot-agent

Figures & tables

Method	No Training	No Prior Knowledge	No Simulation
Ours	$\surd$	$\surd$	$\surd$
ZeroCAP [ 7 ]	$\surd$	$\times$	$\times$
TaskExp [ 6 ]	$\times$	$\surd$	$\times$
SIGMA [ 9 ]	$\times$	$\times$	$\times$
Hybrid [ 10 ]	$\surd$	$\times$	$\surd$
COHERENT [ 8 ]	$\surd$	$\times$	$\surd$

TABLE I: Comparison Against the Triple Zero Criteria

Variable	Description
$L_{T}$	High-level task target location (natural language)
$L_{W}$	Quadruped waypoint location (natural language)
$C$	Inter-agent interaction context
$I_{B}$	Humanoid’s forward perceptual data
$I_{D}$	Quadruped’s forward perceptual data
$d_{dt}$	Euclidean distance to target

TABLE II: Description of Variables

Fig. 1: System pipeline of the proposed multi-agent collaboration framework. Given observations from both the humanoid and quadruped perspectives, the system performs reasoning to determine whether the target is directly reachable. If the destination is visible, the humanoid plans and executes obstacle-aware navigation. Otherwise, the humanoid delegates exploration to the quadruped, which surveys the environment and provides feedback on path feasibility. Based on this feedback, the humanoid updates its plan and continues navigation. This iterative collaboration enables robust decision-making and efficient task completion in complex real-world scenarios. 2 2 2 Robot icons created by Good Ware, robot dog icons created by Izwar Muis - Flaticon

Fig. 2: Iterative decision-making logic of the humanoid agent. The humanoid evaluates path conditions and adaptively determines whether to proceed directly or request assistance from the quadruped. When path conditions are unfavorable, the humanoid iteratively leverages quadruped-guided exploration to identify beneficial waypoints, refining its navigation plan through feedback. Once a feasible waypoint is established, the humanoid proceeds toward it. In contrast, when path conditions are favorable, the humanoid directly selects a waypoint and navigates to the target without additional exploration. This iterative strategy enables robust and efficient navigation under varying environmental conditions.

	$TIME$ $\downarrow$ (s)	$D$ $\downarrow$ (m)	$R$ $\downarrow$ (rad)	$TIME$ $\downarrow$ (s)	$D$ $\downarrow$ (m)	$R$ $\downarrow$ (rad)
	TZPP System			Human baseline
scene 1	64.00	4.00	1.57	53.30	3.80	1.57
scene 2	18.22	2.60	1.05	17.47	2.57	0.79
scene 3	28.58	4.55	2.83	28.01	4.53	2.09
scene 4	120.58	6.80	4.71	80.00	6.47	4.71
scene 5	154.89	14.60	11.78	94.21	13.83	4.29

TABLE III: Task finish comparison of TZPP vs. human baseline

	$PS$ $\uparrow$	$RMSE$ $\downarrow$	$PS$ $\uparrow$	$RMSE$ $\downarrow$
	G1-Go2 System		Human baseline
scene 1	68.18	85.99	82.19	56.05
scene 2	98.08	14.42	99.23	6.24
scene 3	96.74	40.75	97.13	21.21
scene 4	92.35	120.91	97.06	58.52
scene 5	88.18	56.54	93.13	33.22

TABLE IV: Task planning comparison of G1-Go2 system vs. human baseline

	$N_{K}$ $\uparrow$	$EER$ $\uparrow$	$N_{K}$ $\uparrow$	$EER$ $\uparrow$
	G1-Go2 System		Human baseline
scene 1	2	100%	2	100%
scene 2	2	100%	2	100%
scene 3	2	100%	2	100%
scene 4	4	80%	2	100%
scene 5	6	100%	6	100%

TABLE V: Quadrupedal Agent exploration comparison of G1-Go2 system vs. human baseline

	G1-Go2 System			Human baseline
	$N_{E}$	$V_{GE}$ $\downarrow$	$N_{\mathrm{move}}$ $\downarrow$	$N_{E}$	$V_{GE}$ $\downarrow$	$N_{\mathrm{move}}$ $\downarrow$
scene 1	2	0.98	5	2	0.98	2
scene 2	1	0.86	2	1	0.91	2
scene 3	2	0.76	5	1	0.91	2
scene 4	5	0.72	6	4	0.88	5
scene 5	6	0.96	9	5	1.03	6

TABLE VI: Collaboration comparison of G1-Go2 system vs. human baseline

	TIME $\downarrow$	PS $\uparrow$	CR $\uparrow$	RSME $\downarrow$	TIME $\downarrow$	PS $\uparrow$	CR $\uparrow$	RSME $\downarrow$
	G1-Go2				G1-only
scene 1	64.00	68.18	100%	85.99	43.12	67.47	100%	86.69
scene 2	18.22	98.08	100%	14.42	16.14	75.59	100%	124.12
scene 3	28.58	96.74	100%	40.75	N/A	N/A	33.33%	N/A
scene 4	120.58	92.35	100%	120.91	N/A	N/A	40.00%	N/A

TABLE VII: Comparative Ablation Study on Heterogeneous Agents

	Time $\downarrow$	$CCR_{q}$ $\uparrow$	$N_{1}$ $\downarrow$	$N_{2}$ $\downarrow$	$N_{3}$ $\downarrow$
G1(no X)-Go2	43.78	56.27%	0.6	1.4	3
G1-Go2	46.28	86.00%	0	0.2	0.6
Human	54.60	80.00%	0	0	0.2

TABLE VIII: Open area comparison of G1-Go2(without Mode X), G1-Go2 system and human baseline

	Time $\downarrow$	$V_{\mathrm{avoid}}$
G1(no Y)-Go2 System	N/A	0.23
G1-Go2 System	154.89	1.00
Human baseline	94.21	1.00

TABLE IX: Obstacle Handling comparison of G1-Go2(without Mode Y), G1-Go2 system and human baseline

Explore similar work

DeCoNav: Dialog enhanced Long-Horizon Collaborative Vision-Language Navigation
Apr 14, 2026Sunyao Zhou, Yunzi Wu, Tianhang Wang +5Multi-Agent Coordination Vision-Language Navigation
Institute of Artificial Intelligence (TeleAI), China Telecom · Fudan University · Tongji University
DM $^3$ -Nav: Decentralized Multi-Agent Multimodal Multi-Object Semantic Navigation
Apr 23, 2026Amin Kashiri, Atharva Jamsandekar, Yasin YazıcıoğluVision-Language Navigation Multi-Robot Coordination
Northeastern University, Boston, USA · Department of Electrical and Computer Engineering at Northeastern University, Boston, MA
Scale-Plan: Scalable Language-Enabled Task Planning for Heterogeneous Multi-Robot Teams
Mar 9, 2026Piyush Gupta, Sangjae Bae, Jiachen Li +1Large Language Model-Based Robot Planning Symbolic Planning
Honda Research Institute, San Jose, CA 95134, USA. · Georgia Institute of Technology, Atlanta, GA 30332, USA.

Apr 14, 2026J/K move · Enter open · S save

DeCoNav: Dialog enhanced Long-Horizon Collaborative Vision-Language Navigation

Sunyao Zhou, Yunzi Wu, Tianhang Wang, Xinhai Li+4

Institute of Artificial Intelligence (TeleAI), China Telecom · Fudan University · Tongji University

Multi-Agent Coordination Vision-Language Navigation Multi-Robot Coordination Multi-Agent Planning Multi-Robot Task Allocation

Open paper AskPDF arXiv

Long-horizon collaborative vision-language navigation (VLN) is critical for multi-robot systems to accomplish complex tasks beyond the capability of a single agent. CoNavBench takes a first step by introducing the first collaborative long-horizon VLN benchmark with relay-style multi-robot tasks, a collaboration taxonomy, along with graph-grounded generation and evaluation to model handoffs and rendezvous in shared environments. However, existing benchmarks and evaluations often do not enforce strictly synchronized dual-robot rollout on a shared world timeline, and they typically rely on static coordination policies that cannot adapt when new cross-agent evidence emerges. We present Dialog enhanced Long-Horizon Collaborative Vision-Language Navigation (DeCoNav), a decentralized framework that couples event-triggered dialogue with dynamic task allocation and replanning for real-time, adaptive coordination. In DeCoNav, robots exchange compact semantic states via dialogue without a central controller. When informative events such as new evidence, uncertainty, or conflicts arise, dialogue is triggered to dynamically reassign subgoals and replan under synchronized execution. Implemented in DeCoNavBench with 1,213 tasks across 176 HM3D scenes, DeCoNav improves the both-success rate (BSR) by 69.2%, demonstrating the effectiveness of dialogue-driven, dynamically reallocated planning for multi-robot collaboration.

Abstract

Explore similar work

DeCoNav: Dialog enhanced Long-Horizon Collaborative Vision-Language Navigation

DM3^33-Nav: Decentralized Multi-Agent Multimodal Multi-Object Semantic Navigation

Scale-Plan: Scalable Language-Enabled Task Planning for Heterogeneous Multi-Robot Teams

DM $^3$ -Nav: Decentralized Multi-Agent Multimodal Multi-Object Semantic Navigation