Current state of the multi-agent multi-view experimental and digital twin rendezvous (MMEDR-Autonomous) framework

Logan Banker; Michael Wozniak; Mohanad Alameer; Smriti Nandan Paul; David Meisinger; Grant Baer; Trevor Hunting; Ryan Dunham

cs.ROMar 21, 2026

Current state of the multi-agent multi-view experimental and digital twin rendezvous (MMEDR-Autonomous) framework

Authors: Logan Banker, Michael Wozniak, Mohanad Alameer, Smriti Nandan Paul, David Meisinger, Grant Baer, Trevor Hunting, Ryan Dunham,

Organizations: Department of Mechanical and Aerospace Engineering, Missouri University of Science and Technology, Rolla, 65409, MO, United States · Purdue University, 610 Purdue Mall, West Lafayette, 47907, USA

Abstract

As near-Earth resident space objects proliferate, there is an increasing demand for reliable technologies in applications of on-orbit servicing, debris removal, and orbit modification. Rendezvous and docking are critical mission phases for such applications and can benefit from greater autonomy to reduce operational complexity and human workload. Machine learning-based methods can be integrated within the guidance, navigation, and control (GNC) architecture to design a robust rendezvous and docking framework. In this work, the Multi-Agent Multi-View Experimental and Digital Twin Rendezvous (MMEDR-Autonomous) is introduced as a unified framework comprising a learning-based optical navigation network, a reinforcement learning-based guidance approach under ongoing development, and a hardware-in-the-loop testbed. Navigation employs a lightweight monocular pose estimation network with multi-scale feature fusion, trained on realistic image augmentations to mitigate domain shift. The guidance component is examined with emphasis on learning stability, reward design, and systematic hyperparameter tuning under mission-relevant constraints. Prior Control Barrier Function results for Clohessy-Wiltshire dynamics are reviewed as a basis for enforcing safety and operational constraints and for guiding future nonlinear controller design within the MMEDR-Autonomous framework. The MMEDR-Autonomous framework is currently progressing toward integrated experimental validation in multi-agent rendezvous scenarios.

Figures & tables

Figure 1: Chaser and Target Spacecraft in LVLH Frame, Figure Adapted from Khoury (2020)

Figure 2: Overall CNN Architecture for Optical Navigation, Based on Combination of MobileNetV3Large, FPN, and Deep-6DPose ( Howard et al., 2019 ; Lin et al., 2017 ; Do et al., 2018 )

Augment Type	Probability - Lightbox	Probability - Sunlamp
Brightness and Contrast	0.8	0.8
Gaussian Noise	0.7	0.8
Motion Blur	0.7	0.8
ISO Noise	0.7	0.8
Gamma	0.6	0.7
Sun Flare	-	0.7

Table 1: List of Augmentations and Probability of Application

Figure 3: Current State of the DCV-Space Testing Facility

Link	$\theta$	$d$ (mm)	$a$ (mm)	$\alpha$ (deg)
1	$\theta_{1}$	230	0	90
2	$\theta_{2}$	0	825.2	0
3	$\theta_{3}$	0	746	0
4	$\theta_{4}$	175.6	0	-90
5	$\theta_{5}$	128.8	0	90
6	$\theta_{6}$	136.5	0	0

Table 2: D-H Parameters for Dobot CR20A Robot

Figure 5: Overview of All MMEDR Subsystems

Panel of Figure 6 — Figure 6: Docking Results from Manual Shaping and a 50m Initial Distance

Panel of Figure 7 — Figure 7: Successful Iterations of Bayesian optimization

Figure 8: Bayesian optimization Trajectory with Velocity Constraints

Training	Lightbox			Sunlamp
Augments	$E_{t}$ (m)	$E_{q}$ ( ∘ )	$E_{p}^{*}$ (-)	$E_{t}$ (m)	$E_{q}$ ( ∘ )	$E_{p}^{*}$ (-)
Without	1.95	76.25	1.65	1.93	89.59	1.85
With	0.36	45.65	0.86	0.52	66.74	1.28

Table 3: Testing Results for Lightbox and Sunlamp Domains, With and Without Data Augmentation (Described in Section 4.2.1 )

Model	Lightbox			Sunlamp
Model	$E_{t}$ (m)	$E_{q}$ ( ∘ )	$E_{p}^{*}$ (-)	$E_{t}$ (m)	$E_{q}$ ( ∘ )	$E_{p}^{*}$ (-)
SPN	0.45	65.12	1.21	0.65	92.95	1.73
KRN	2.25	44.53	1.12	14.64	80.95	3.73
Ours	0.36	45.65	0.86	0.52	66.74	1.28

Table 4: Comparison of Synthetic-Only Training, Tested on Lightbox and Sunlamp Domains

Panel of Figure 9 — Figure 9: Synthetic Images of the Aura Spacecraft Generated in Blender

Spacecraft Rendezvous and Proximity Operations

Spacecraft rendezvous and proximity operations cover the navigation, guidance, and control needed to approach, inspect, dock with, or capture spacecraft safely.

See all 26 papers

From Orbit Matching to Docking: A Hierarchical Reinforcement-Learning Approach to Multiphase Rendezvous Trajectory Design
Oct 12, 2026Mariko Storey-Matsutani, Anatoly Zlotnik, Lee Holguin +2Spacecraft Rendezvous and Proximity Operations RL Control
Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA · Los Alamos National Laboratory, Los Alamos, NM, USA
Learning Adaptive Multi-Task Guidance, Navigation, and Control via Hypernetworks
Jul 27, 2026Ricard Marsal I Castan, Aman Arora, Antoine Richard +3Robotic Control Hypernetworks
University of Luxembourg · 2Georgia Institute of Technology
AI-Augmented Model Predictive Control for Safe and Adaptive Rendezvous and Proximity Operations
Jul 18, 2026Luca Sportelli, Tyler Barr, Cagri Kilic +1Spacecraft Rendezvous and Proximity Operations Multi-Agent Control
Department of Aerospace Engineering, Embry–Riddle Aeronautical University, Daytona Beach, FL 32114

Oct 12, 2026J/K move · Enter open · S save

From Orbit Matching to Docking: A Hierarchical Reinforcement-Learning Approach to Multiphase Rendezvous Trajectory Design

Mariko Storey-Matsutani, Anatoly Zlotnik, Lee Holguin, Zigfried Hampel-Arias+1

Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA · Los Alamos National Laboratory, Los Alamos, NM, USA

Spacecraft Rendezvous and Proximity Operations RL Control Hierarchical RL

Open paper AskPDF arXiv

As the complexity of space missions increases, autonomous spacecraft guidance must accommodate transitions between mission phases with distinct dynamics, objectives, and constraints. Rendezvous and proximity operations (RPO) are particularly challenging because they span multiple operating regimes, from orbital-scale matching and far-range approach to close-range maneuvering and terminal docking. Existing reinforcement-learning approaches generally address only a subset of this sequence or rely on manually defined switching conditions between controllers. This study presents a hierarchical reinforcement-learning framework for autonomous, three-degree-of-freedom RPO that spans the complete trajectory from orbit matching to docking. Five specialized policies are trained for orbit matching, far-range approach, mid-range approach, close-range approach, and docking. A learned high-level manager policy then selects among these phase-specific controllers based on the evolving spacecraft state. The hierarchical approach is evaluated over the full end-to-end rendezvous sequence and compared with a fixed sequential-invocation heuristic to assess trajectory performance and transition behavior. Training and inference metrics are also analyzed for the individual phase policies and the complete hierarchical system.

Abstract

Spacecraft Rendezvous and Proximity Operations

Explore similar work

From Orbit Matching to Docking: A Hierarchical Reinforcement-Learning Approach to Multiphase Rendezvous Trajectory Design

Learning Adaptive Multi-Task Guidance, Navigation, and Control via Hypernetworks

AI-Augmented Model Predictive Control for Safe and Adaptive Rendezvous and Proximity Operations