Research

The MICO Lab develops control, optimization, and learning algorithms across three levels: individual robot capabilities, adaptation and interaction between agents, and coordination across multi-robot and networked systems.

Representative papers are grouped by research direction below. A complete list is available on the Publication page and Google Scholar.

Transfer Learning and Agent Adaptation

We study how learned policies transfer across related tasks and adapt to changes in other agents, combining context-aware robot learning with policy sensitivity.

Policy adaptation to changes in another agent through finite-depth sensitivity estimation

Adapting to Changes in Agent Behavior via Finite-Depth Policy Sensitivity

Lan Shi, Daigo Shishika, and Xuan Wang. arXiv, 2026.

Estimates policy sensitivity at adjustable propagation depths to adapt to another agent's behavioral changes, improving zero-shot transfer and subsequent fine-tuning in a pursuit-evasion game.

Paper

Context-aware robot learning figure

CARoL: Context-Aware Adaptation for Robot Learning

Zechen Hu, Tong Xu, Xuesu Xiao, and Xuan Wang. IEEE Robotics and Automation Letters, 2025.

Context-aware reinforcement learning for efficient adaptation across related robot tasks.

Paper

Legged Mobile Manipulation and Motion Planning

We develop contact-aware loco-manipulation and geometry-aware motion planning to help individual robots perform demanding tasks in constrained environments.

Legged mobile manipulator using environmental contacts for physical support

Contact as a Decision Variable: Capability-Tradeoff Contact Selection for Legged Loco-Manipulation

Al Jaber Mahmud, Shuai Li, and Xuan Wang. arXiv, 2026.

Selects environmental support contacts by balancing residual wrench capability, end-effector reach, and base mobility, with efficient screening and sensitivity-based evaluation.

Paper

Geometry-aligned LLM planning figure

Geometry-Aligned LLM Fine-Tuning for Sequential Narrow-Opening Planning

Al Jaber Mahmud, and Xuan Wang. arXiv.

This paper proposes a geometry-aligned LLM fine-tuning framework for rigid-body motion planning through multiple sequential narrow openings. The method generates fixed-length, machine-readable waypoint sequences that are geometrically feasible and coordinated across openings.

Paper

Elastic Models for Mobile Robots

We build sequence models that can be trained once and resized for deployment across devices with different memory, computing, and latency budgets.

Elastic Selective Spectral Hybrid architecture and adjustable model capacities

Elastic Selective Spectral Hybrids for Train-Once, Export-Many Budgeted Inference

Dachuan Song, Chuchu Chen, and Xuan Wang. arXiv, 2026.

Combines input-dependent spectral recurrence with sliding-window attention to export compact models from a single training run for different computing and latency budgets.

Paper

Figure 1 from the ES-SSM paper: residual block and spectral sublayer with budgeted channel gates

Elastic Spectral State Space Models for Train-Once Budgeted Inference

Dachuan Song, Junyu Yin, Zechen Hu, and Xuan Wang. NeurIPS, 2026.

Uses ordered spectral channels, adaptive gates, and budget dropout to train once and export smaller sequence models across a range of resource budgets.

Paper GitHub

Human-Robot Interaction and Collaborative Autonomy

We develop control and planning methods for robots that coordinate with humans, adapt to human uncertainty, and provide safety-aware assistance in collaborative manipulation and transportation tasks.

Human-robot co-transportation paper preview

Human-Robot Co-Transportation using Disturbance-Aware MPC with Pose Optimization

Al Jaber Mahmud, Amir Hossain Raj, Duc M. Nguyen, Weizi Li, Xuesu Xiao, and Xuan Wang. IEEE/RSJ IROS, 2025.

Whole-body control for human-robot co-transportation under disturbance and human uncertainty. Prelimiary work got Best Paper Finalist at the ICRA Workshop on Exploring Role Allocation in Human-Robot Co-Manipulation.

Paper Supplement

Human-robot co-transportation mutual adaptation figure

Mutual Adaptation in Human-Robot Co-Transportation with Human Preference Uncertainty

Al Jaber Mahmud, Weizi Li, and Xuan Wang. arXiv.

It models probabilistic human choices, introduces a time-varying stubbornness measure for coordination-mode transitions, and uses pose optimization to improve task performance when adapting to human behavior.

Paper

Distributed resource allocation for human-autonomy teaming figure

Distributed Resource Allocation for Human-Autonomy Teaming With Human Preference Uncertainty

Yichen Yao, Ryan Mbagna Nanko, Yue Wang, and Xuan Wang. IEEE Control Systems Letters.

It develops a distributed optimization framework that incorporates human response models and enables autonomous agents to coordinate resource decisions while accounting for preference uncertainty in human teammates.

Paper GitHub

Multi-Robot Coordination, Reinforcement Learning, and Games

We study scalable coordination for teams of robots operating under local communication, coupled objectives, adversarial uncertainty, and graph-structured tasks. The work combines distributed optimization, game theory, learning from demonstrations, and planning on graphs.

Robot learning coordination figure

Bi-CL: A Reinforcement Learning Framework for Robots Coordination Through Bi-level Optimization

Zechen Hu, Daigo Shishika, Xuesu Xiao, and Xuan Wang. IEEE/RSJ IROS, 2024.

A reinforcement learning framework that connects coordination learning with bi-level optimization.

Paper

D3G paper preview

D3G: Learning Multi-Robot Coordination from Demonstrations

Yizhi Zhou, Wanxin Jin, and Xuan Wang. IEEE/RSJ IROS, 2024.

Distributed differentiable dynamic games for learning coordination strategies from demonstrations.

Paper Video

Robot learning coordination figure

Learning Coordinated Maneuver in Adversarial Environments

Zechen Hu, Manshi Limbu, Daigo Shishika, Xuesu Xiao, and Xuan Wang. IEEE/RSJ IROS, 2024.

A reinforcement learning framework that enables robot coordination in adversarial environments.

Paper

Team coordination paper preview

Team Coordination on Graphs with State-Dependent Edge Costs

Manshi Limbu, Zechen Hu, Sara Oughourli, Xuan Wang, Xuesu Xiao, and Daigo Shishika. IEEE/RSJ IROS, 2023.

Best Paper Award on Cognitive Robotics finalist.

Paper Video

Team coordination graph example

Team Coordination on Graphs: Problem, Analysis, and Algorithms

Yanlin Zhou, Manshi Limbu, Gregory J. Stein, Xuan Wang, Daigo Shishika, and Xuesu Xiao. IEEE/RSJ IROS, 2024.

Graph-based coordination analysis and algorithms for multi-robot teams.

Paper

Adversarial_Graph

Multi-Robot Coordination in an Adversarial Graph-Traversal Game

James Berneburg, Xuan Wang, Xuesu Xiao, and Daigo Shishika. IEEE/RSJ IROS, 2025.

Coordination under adversarial graph traversal and strategic task constraints.

Paper

Distributed Optimization, Control, and Resilient Computing

We develop distributed algorithms for optimization, consensus, linear equation solving, information fusion, and resilient decision making in networked systems with communication constraints and adversarial failures.

Resilient consensus figure

Resilience for Distributed Consensus with Constraints

Xuan Wang, Shaoshuai Mou, and Shreyas Sundaram. IEEE Transactions on Automatic Control, accepted, 2025.

Resilient consensus algorithms for constrained distributed networks under hostile behavior.

Paper

Distributed optimization figure

Constrained Consensus-Based Distributed Optimization with Integral Feedback

Xuan Wang, Shaoshuai Mou, and B. D. O. Anderson. IEEE Transactions on Automatic Control, 2022.

Integral-feedback distributed optimization with reduced communication and exponential convergence.

Paper

Double-layered network figure

Scalable, Distributed Algorithms for Solving Linear Equations via Double-Layered Networks

Xuan Wang, Shaoshuai Mou, and B. D. O. Anderson. IEEE Transactions on Automatic Control, 2020.

Double-layered network structure for scalable distributed linear-equation solving.

Paper

Distributed task allocation figure

Distributed Algorithm with Resilience for Multi-Agent Task Allocation

Xuan Wang, Jeffrey Hudack, and Shaoshuai Mou. IEEE ICPS, 2021.

Best Paper Award. Auction-consensus task allocation with resilience to temporal attacks.

Paper

Data-Driven Control of Networked Systems

Beyond robot teams, we use data-driven control and reconstruction methods to understand networked dynamics, including neural mass models and fMRI-based system identification.

fMRI fingerprint paper preview

Reconstructing brain causal dynamics for subject and task fingerprints using fMRI time-series data

Dachuan Song, Li Shen, Duy Duong-Tran, and Xuan Wang. Health Information Science and Systems.

Dynamics reconstruction for subject and task identification from fMRI time-series data.

Paper Video

Brain network reconstruction figure

Efficient Reconstruction of Neural Mass Dynamics Modeled by Linear-Threshold Networks

Xuan Wang and Jorge Cortes. IEEE Transactions on Automatic Control, 2025.

Data-driven reconstruction of network dynamics from observed time-series behavior.

Paper

Data-driven control figure

Data-Driven Control of Linear-Threshold Network Dynamics

Xuan Wang and Jorge Cortes. American Control Conference, 2022.

Model-free control methods for linear-threshold network dynamics.

Paper

Cooperative SLAM, Localization, and State Estimation

We design consistent state-estimation methods for multi-robot localization and navigation, combining visual-inertial sensing, UWB/ranging, Lie group structure, and distributed cooperative estimation.

Distributed 3-D multi-robot cooperative localization figure

Distributed 3-D Multi-Robot Cooperative Localization: An Efficient and Consistent Approach

Yizhi Zhou, Yufan Liu, and Xuan Wang. IEEE Robotics and Automation Letters.

This paper proposes a distributed cooperative localization method for multi-robot systems in 3-D environments. By performing information fusion directly on Lie groups, the method accounts for nonlinear 3-D rotations and translations while preserving observability consistency in distributed localization.

Paper Supplement

CVIRO paper preview

CVIRO: A Consistent and Tightly-Coupled Visual-Inertial-Ranging Odometry on Lie Groups

Yizhi Zhou, Zhe Kang, Junyi Xia, and Xuan Wang. IEEE/RSJ IROS, 2025.

Tightly coupled visual-inertial-ranging odometry with consistency on Lie groups.

Paper Supplement

UWB-aided visual-inertial navigation paper preview

Robust Online Calibration for UWB-Aided Visual-Inertial Navigation with Bias Correction

Yizhi Zhou, Jinle Xu, Junyi Xia, Zechen Hu, Weizi Li, and Xuan Wang. IEEE/RSJ IROS, 2025.

Robust online calibration and bias correction for UWB-aided visual-inertial systems.

Paper Supplement