Cross-Robot Behavior Adaptation through Intention Alignment

Yuan; Chen, Xi

doi:10.5281/zenodo.18618978

Published February 12, 2026 | Version v1

Software Open

Cross-Robot Behavior Adaptation through Intention Alignment

Title:
Cross-Robot Behavior Adaptation through Intention Alignment - Experimental Data and Analysis Scripts

Authors:
Xi Chen*, Yuan Gao*, Hangxin Liu, Fangkai Yang, Ali Ghadirzadeh, Jun Yang, Bin Liang, Chongjie Zhang, Tin Lun Lam, Song-Chun Zhu

Description:

This repository contains the experimental data and analysis scripts used in the paper "Cross-Robot Behavior Adaptation through Intention Alignment"

Repository Contents:

realrobot-result/ - Real-world experiment data

real-result.xlsx: Records of 30 real-world imitation scenarios, including:
- Motions performed by the three demonstrator robots (Tello, Dual-Arm, Spark)
- Scenario types
- Learner robots presented in each scenario (Cuboat, Diablo, Single-Arm, Pepper)
- Motions performed by the learner robots given the demonstration
- Imitation results (success/outcome)
Used to construct Table 1 and Table 2 in the paper

latent-result/ - Latent intention-space analysis data

Latent embeddings of 120 samples used to create Table 3
Analyzes the structure of the learned intention space:
- Intra-class distance
- Cross-embodiment error
- Global inter-class distance
Analysis script: print_latent_analysis.py

simulation-result/ - Simulation experiment results

Evaluation results for two simulation tasks: monitoring and item picking
Pre-computed results for statistical comparisons against baseline methods
Used to construct Tables S3–S6
Analysis script: print_simulation_stats.py

Reproduction Instructions:

bash

# Table 3 (latent space analysis)
python print_latent_analysis.py

# Tables S3–S6 (simulation statistics)
python print_simulation_stats.py --task navigation --method usc    # Table S3
python print_simulation_stats.py --task navigation --method language  # Table S4
python print_simulation_stats.py --task mp1 --method usc          # Table S5
python print_simulation_stats.py --task mp1 --method language     # Table S6

Related Publication:

Chen, X.*, Gao, Y.*, Liu, H., Yang, F., Ghadirzadeh, A., Yang, J., Liang, B., Zhang, C., Lam, T. L., & Zhu, S. C. (2026). Cross-Robot Behavior Adaptation through Intention Alignment.

Funding:

This work was supported by the National Key R&D Program of China (Grant Nos. 2024YFB4505500 and 2024YFB4505503), National Natural Science Foundation of China (Grant No. 62376031), Shenzhen Science and Technology Program (Grant No. JSGGKQTD20221101115656029 and ZDCY20250901104706008), and Guangdong Basic and Applied Basic Research Foundation (Grant No. 2023B1515020089).

Keywords:
imitation learning, heterogeneous robots, intention alignment, multi-robot systems, cross-embodiment transfer, behavior adaptation

Related identifiers:

GitHub repository: https://github.com/pcchenxi/Cross-Robot-Behavior-Adaptation-through-Intention-Alignment

Version: v1.0.0

Files