CableRobotGraphSim: A Graph Neural Network for Modeling Partially Observable Cable-Driven Robot Dynamics

Nelson Chen; William R. Johnson; Rebecca Kramer-Bottiglio; Kostas Bekris; Mridul Aanjaneya

cs.ROFeb 24, 2026

CableRobotGraphSim: A Graph Neural Network for Modeling Partially Observable Cable-Driven Robot Dynamics

Authors: Nelson Chen, William R. Johnson, Rebecca Kramer-Bottiglio, Kostas Bekris, Mridul Aanjaneya

Organizations: 1Rutgers University · 2Yale University

Abstract

General-purpose simulators have accelerated the development of robots. Traditional simulators based on first-principles, however, typically require full-state observability or depend on parameter search for system identification. This work presents CableRobotGraphSim, a novel Graph Neural Network (GNN) model for cable-driven robots that aims to address shortcomings of prior simulation solutions. By representing cable-driven robots as graphs, with the rigid-bodies as nodes and the cables and contacts as edges, this model can quickly and accurately match the properties of other simulation models and real robots, while ingesting only partially observable inputs. Accompanying the GNN model is a sim-and-real co-training procedure that promotes generalization and robustness to noisy real data. This model is further integrated with a Model Predictive Path Integral (MPPI) controller for closed-loop navigation, which showcases the model's speed and accuracy.

Figures & tables

Figure 1: Left: Static, open-source 3-bar tensegrity platform. Right: The platform rolling clockwise.

Figure 2: The three stages of a simulation step. Left : Graph and feature generation, where state

\mathbf{X}_{t}

, rest length

\ell_{t}^{rest}

, control history

U_{t-h:t}

, and future controls

U_{t+1:t+n}

form the graph

\mathcal{G}_{t}=(\mathcal{V}_{t},\mathcal{E}_{t})

. Middle:

\mathcal{G}_{t}

is passed through MLP encoders and an LSTM block (with hidden state

H_{t}

and memory

C_{t}

) to produce

H_{t+1},C_{t+1}

H_{t+1}

then undergoes

L

rounds of message passing, after which the latent node and cable-edge vectors are decoded into

n

velocity and rest-length changes. Right: The predicted chunk is integrated iteratively to advance

\ell_{t}^{rest}

\ell_{t+n}^{rest}

and

\mathbf{X}_{t}

\mathbf{X}_{t+n}

Figure 3: Sim-and-real co-training. Multiple simulation datasets are generated by varying system parameters, and are pooled with a real dataset of unknown system parameters. Data points are distinguished by a one-hot encoding node feature of the dataset enumerated index.

Figure 4: Sim2sim evaluation for full trajectories and short horizons, as well as a 3-bar and 6-bar tensegrity.

Model	Pos Error (%)	Rot Error (°)	Pos Error (%)	Rot Error (°)
	Full Traj	Full Traj	Short Horizon	Short Horizon
Hybrid GNN - uninformed	Unstable	Unstable	Unstable	Unstable
Hybrid GNN - heuristic	91.12	35.29	N/A	N/A
T-DPE - uninformed	465.59	60.32	9.59	28.76
T-DPE - heuristic	102.76	38.48	N/A	N/A
CableRobotGraphSim (ours)	48.95	26.53	3.95	7.031

Table 1: Real2sim evaluation results on a real 3-bar tensegrity platform.

Figure 5: Running MPPI controller in a loop of model training, task execution, and new data generation for model retraining over three iterations. Positional and rotational errors are evaluated on a pooled test set from all iterations. Success rate and task completion time are measured per iteration.

Figure 6: S-shaped obstacle course for navigating from the red to the green sphere.

Figure 7: Maze map for navigating from the red to the green sphere.

	Full Traj	Full Traj	Short Horizon	Short Horizon
Model	Pos Error (%)	Rot Error (°)	Pos Error (%)	Rot Error (°)
(baseline) Hybrid GNN	Unstable	Unstable	Unstable	Unstable
(i) Fully Learned GNN (F-GNN)	$20.6\pm 26.14$	$47.1\pm 27.03$	$0.11\pm 0.04$	$4.91\pm 0.99$
(ii) F-GNN + Config Feats (FC-GNN)	$1.15\pm 1.13$	$15.43\pm 15.54$	$0.1\pm 0.04$	$4.6\pm 1.21$
(iii) FC-GNN + Recurrent (FCR-GNN)	$\mathbf{1.04\pm 1.3}$	$\mathbf{7.61\pm 4.72}$	$\mathbf{0.06\pm 0.02}$	$\mathbf{3.24\pm 0.88}$
FCR-GNN + analytical motor	$4.49\pm 2.21$	$31.18\pm 7.22$	$0.1\pm 0.04$	$5.14\pm 0.66$

Table 2: GNN architecture component ablation study. The first four rows compare GNN components using ground-truth motor outputs. The last two rows compare an analytical vs. a learned motor model.

Figure 8: Sim-and-real co-training ablation study for varying number of simulation datasets.

Figure 9: Prediction chunk size ablation over the number of forward time steps per simulation call.

Robotics Simulation

Robotics simulation builds virtual robots and environments to model, train, and evaluate autonomous systems before real-world deployment.

See all 320 papers

SILO: Simulation-in-the-Loop Sim-to-Real Transfer for Multi-Stage Cable Routing
Jul 6, 2026Stone Tao, Jie Xu, Hesam Rabeti +3RL for Robotics Sim-to-Real Transfer
1NVIDIA Corporation · University of California, San Diego
Model Predictive Control of Tensegrity Robots via Contact-Aware Graph Neural Dynamics Model
Sep 8, 2026Nelson Chen, Patrick Meng, Charles Tang +5Robotic Control Robot Dynamics Identification
Computer Science, Rutgers University, Piscataway, NJ, USA. · Mechanical Engineering, Yale University, New Haven, CT, USA.
CableVLA: Simulation-Privileged Global-Local Representation Learning for Cable Routing
Sep 22, 2026Zhifei Teng, Bo Feng, Xiang Zou +4Visuo-Tactile Representation Learning Sim-to-Real Transfer
State Key Laboratory of Intelligent Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan, China. · School of Mathematics, Harbin Institute of Technology, Harbin, China.

Jul 6, 2026J/K move · Enter open · S save

SILO: Simulation-in-the-Loop Sim-to-Real Transfer for Multi-Stage Cable Routing

Stone Tao, Jie Xu, Hesam Rabeti, Yashraj Narang+2

1NVIDIA Corporation · University of California, San Diego

RL for Robotics Sim-to-Real Transfer Deformable Object Manipulation Deformable Object Simulation

Open paper AskPDF arXiv

Linear-deformable manipulation remains challenging due to the complex deformations of objects such as cables and ropes. Prior data-driven approaches, particularly imitation learning, have shown some promise in narrowly defined settings but typically require thousands of demonstrations for specific tasks and cable types, limiting scalability and generalization. We introduce a sim-to-real reinforcement learning (RL) framework for multi-stage cable routing that leverages GPU-parallelized simulation to approximate linear deformable behaviors. Training across thousands of parallel simulations enables the learned policies to generalize across diverse cable geometries and deformation patterns. To bridge the sim-to-real gap, we propose a novel deployment strategy that combines a Simulation In the LOop (SILO) execution framework, localized RL policies, and robust cable state estimation. On real-world cable routing tasks, our approach achieves higher success rates and 2x reduction in cycle times compared to prior state-of-the-art learning methods. To our knowledge, this is the first successful sim-to-real transfer of RL policies for multi-stage cable routing. Videos and additional visualizations are available at https://silo-cable-routing.github.io/

Abstract

Robotics Simulation

Explore similar work

SILO: Simulation-in-the-Loop Sim-to-Real Transfer for Multi-Stage Cable Routing

Model Predictive Control of Tensegrity Robots via Contact-Aware Graph Neural Dynamics Model

CableVLA: Simulation-Privileged Global-Local Representation Learning for Cable Routing