Research

overview & theory

I study contact-rich dexterous manipulation under uncertainty, connecting ideas from risk-sensitive control and planning, belief estimation, and policy learning to make manipulation with multi-fingered hands more reliable when object pose, friction, contact modes, and target geometry are only partially observed. My published work addresses robust stochastic planning, safe control via distributional reinforcement learning, and multi-agent coordination under uncertainty. Building on these foundations, my current focus is on grasp synthesis, visuo-tactile policy learning, and uncertainty-aware decision-making for dexterous manipulation—studying how robots can estimate uncertainty from proprioceptive, perceptual, and tactile feedback, recognize when a policy is entering a likely failure mode, and decide when to recover, replan, or gather more information. I validate these ideas in MuJoCo, Drake, reinforcement-learning environments, and on real robot platforms equipped with anthropomorphic and tactile-sensorized dexterous hands.

Applications

I apply the aforementioned ideas to deliver solutions to dexterous manipulation problems spanning:

Click any clip to enlarge.

MuJoCo belief-space grasping demo

Robust Dexterous Grasping & Manipulation

Robust visuo-tactile dexterous grasping under uncertain object pose, friction, and contact modes.

Reach-aware grasp planning demo

Reach-Aware SE(3)-Equivariant Grasp Synthesis

Reach-aware, scene-aware, collision-free, and executable multifingered grasp generation.

EquiDexFlow
Policy learning demo

Diffusion Policy Learning from Multimodal Data

Learning from uncertainty-aware expert demonstrations for contact-rich manipulation tasks.

Safe manipulator control demo

Dexterous Data Generation, Data Collection, & Policy Execution

Scene-aware and collision-free VR-assisted data collection and risk-aware execution for learned manipulation policies.

Publications

Preprints (Submitted/Under Review)
2026

arXiV
Clinton Enwerem, John S. Baras, and Calin Belta, “EquiDexFlow: Contact-Grounded SE(3)-Equivariant Dexterous Grasp Generative Flows,” arXiv preprint, 2026. arXiv link
paper project 
@online{enweremEquiDexFlowContactGroundedSE32026,
title = {EquiDexFlow: Contact-Grounded SE(3)-Equivariant Dexterous Grasp Generative Flows},
author = {Enwerem, Clinton and Baras, John S. and Belta, Calin},
year = {2026},
eprint = {2606.12728 [cs.RO]},
eprinttype = {arXiv},
eprintclass = {cs.RO},
doi = {10.48550/arXiv.2606.12728},
url = {http://arxiv.org/abs/2606.12728},
pubstate = {prepublished},
keywords = {Computer Science - Robotics, Computer Science - Machine Learning},
}
arXiV
Clinton Enwerem, John S. Baras, and Calin Belta, “Risk-Constrained Belief-Space Optimization for Safe Control under Latent Uncertainty,” arXiv preprint, 2026. arXiv link
paper 
@online{enweremRiskConstrainedBeliefOptimization2026,
title = {Risk-Constrained Belief-Space Optimization for Safe Control under Latent Uncertainty},
author = {Enwerem, Clinton and Baras, John S. and Belta, Calin},
year = {2026},
eprint = {2604.03868},
eprinttype = {arXiv},
eprintclass = {cs.RO},
doi = {10.48550/arXiv.2604.03868},
url = {http://arxiv.org/abs/2604.03868},
urldate = {2026-04-07},
pubstate = {prepublished},
keywords = {Computer Science - Robotics, Electrical Engineering and Systems Science - Systems and Control},
}
Peer-Reviewed
2026

IROS
Clinton Enwerem, Shreya Kalyanaraman, John S. Baras, and Calin Belta, “Variational Neural Belief Parameterizations for Robust Dexterous Grasping under Multimodal Uncertainty,” 2026. To appear in the Proceedings of the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). arXiv link
paper 
@inproceedings{enweremVariationalNeuralBeliefParameterizations2026,
title = {Variational Neural Belief Parameterizations for Robust Dexterous Grasping under Multimodal Uncertainty},
author = {Enwerem, Clinton and Kalyanaraman, Shreya and Baras, John S. and Belta, Calin},
year = {2026},
booktitle = {Proceedings of the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
eprint = {2604.25897},
eprinttype = {arXiv},
eprintclass = {cs.RO},
doi = {10.48550/arXiv.2604.25897},
url = {http://arxiv.org/abs/2604.25897},
urldate = {2026-04-28},
pubstate = {forthcoming},
keywords = {Computer Science - Robotics, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Systems and Control},
}
2025

CDC
Clinton Enwerem, Aniruddh G. Puranic, John S. Baras, and Calin Belta, Safety-Aware Reinforcement Learning for Control via Risk-Sensitive Value Iteration and Quantile Regression. In the proceedings of the 64th IEEE Conference on Decision and Control (CDC), 2025.
paper 
@online{enweremSafetyAwareRLforControl2025,
title = {Safety-Aware Reinforcement Learning for Control via Risk-Sensitive Value Iteration and Quantile Regression},
author = {Enwerem, Clinton and Puranic, Aniruddh G. and Baras, John S. and Belta, Calin},
year = {2025},
eprint = {2506.06954},
eprinttype = {arXiv},
eprintclass = {cs.LG},
doi = {10.48550/arXiv.2506.06954},
url = {http://arxiv.org/abs/2506.06954},
urldate = {2025-06-08},
pubstate = {prepublished},
keywords = {Computer Science - Machine Learning,Computer Science - Robotics},
}
2024

CDC
Clinton Enwerem, Erfaun Noorani, John S. Baras, and Brian M. Sadler, Robust Stochastic Shortest-Path Planning via Risk-Sensitive Incremental Sampling, In the proceedings of the 63rd IEEE Conference on Decision and Control (CDC), 2024.
paper 
@online{enweremRobustStochasticShortestPath2024b,
title = {Robust {{Stochastic Shortest-Path Planning}} via {{Risk-Sensitive Incremental Sampling}}},
author = {Enwerem, Clinton and Noorani, Erfaun and Baras, John S. and Sadler, Brian M.},
year = {2024},
eprint = {2408.08668},
eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2408.08668},
url = {http://arxiv.org/abs/2408.08668},
urldate = {2024-12-20},
pubstate = {prepublished},
keywords = {Computer Science - Artificial Intelligence,Computer Science - Robotics,Computer Science - Systems and Control,Electrical Engineering and Systems Science - Systems and Control},
}
ECC
Clinton Enwerem and John S Baras. Safe Collective Control under Noisy Inputs and Competing Constraints via Non-Smooth Barrier Functions. In the 2024 European Control Conference (ECC), pp. 3762–3768. IEEE, 2024.
paper 
@inproceedings{enweremSafeCollectiveControl2024,
title = {Safe {{Collective Control Under Noisy Inputs}} and {{Competing Constraints}} via {{Non-Smooth Barrier Functions}}},
booktitle = {2024 {{European Control Conference}} ({{ECC}})},
author = {Enwerem, Clinton and Baras, John S.},
year = {2024},
pages = {3762--3768},
doi = {10.23919/ECC64448.2024.10591027},
url = {https://ieeexplore.ieee.org/abstract/document/10591027},
urldate = {2024-12-20},
eventtitle = {2024 {{European Control Conference}} ({{ECC}})},
keywords = {control barrier functions,Control systems,multi-agent systems,Polynomials,Robustness,Safety,safety-critical control,Smoothing methods,stochastic model-predictive control,Stochastic processes,Upper bound}
}
LCSS
Clinton Enwerem and John S. Baras, Formation Tracking for a Class of Uncertain Multiagent Systems: A Distributed Kalman Filtering Approach, IEEE Control Systems Letters, Volume 8, 2024.
paper 
@article{enweremFormationTrackingClass2024a,
title = {Formation {{Tracking}} for a {{Class}} of {{Uncertain Multi-Agent Systems}}: {{A Distributed Kalman Filtering Approach}}},
shorttitle = {Formation {{Tracking}} for a {{Class}} of {{Uncertain Multi-Agent Systems}}},
author = {Enwerem, Clinton and Baras, John S.},
year = {2024},
journaltitle = {IEEE Control Systems Letters},
volume = {8},
pages = {217--222},
issn = {2475-1456},
doi = {10.1109/LCSYS.2024.3364987},
url = {https://ieeexplore.ieee.org/abstract/document/10433084},
urldate = {2024-12-20},
eventtitle = {{{IEEE Control Systems Letters}}},
keywords = {Collective control,distributed Kalman filtering (DKF),Filtering,formation tracking,Kalman filters,leader-follower networks,multi-agent systems,Noise measurement,Robot kinematics,Sensors,Standards,Uncertainty}
}

2023

CoDIT
Clinton Enwerem and John S. Baras, "Consensus-Based Leader-Follower Formation Tracking for Control-Affine Nonlinear Multiagent Systems," in the 9th International Conference on Control, Decision and Information Technologies (CoDIT), Rome, Italy: IEEE, Jul. 2023, pp. 1226–1231. doi: 10.1109/CoDIT58514.2023.10284199.
paper 
@inproceedings{enweremConsensusBasedLeaderFollowerFormation2023a,
title = {Consensus-{{Based Leader-Follower Formation Tracking}} for {{Control-Affine Nonlinear Multiagent Systems}}},
booktitle = {2023 9th {{International Conference}} on {{Control}}, {{Decision}} and {{Information Technologies}} ({{CoDIT}})},
author = {Enwerem, Clinton and Baras, John S.},
year = {2023},
pages = {1226--1231},
issn = {2576-3555},
doi = {10.1109/CoDIT58514.2023.10284199},
url = {https://ieeexplore.ieee.org/abstract/document/10284199},
urldate = {2024-12-20},
eventtitle = {2023 9th {{International Conference}} on {{Control}}, {{Decision}} and {{Information Technologies}} ({{CoDIT}})},
keywords = {Aerospace electronics,consensus,Control systems,control-affine nonlinear systems,formation control,multiagent systems,Numerical models,Numerical simulation,Topology,Trajectory,Trajectory tracking},
}
Preprints/Reports
2023

arXiV
Clinton Enwerem, John S. Baras, and Danilo Romero, "Distributed Optimal Formation Control for an Uncertain Multiagent System in the Plane," arXiv:2301.05841 [cs, eess], Jan. 2023.
paper 
@online{enweremDistributedOptimalFormation2023,
title = {Distributed {{Optimal Formation Control}} for an {{Uncertain Multiagent System}} in the {{Plane}}},
author = {Enwerem, Clinton and Baras, John and Romero, Danilo},
date = {2023-01-14},
eprint = {2301.05841},
eprinttype = {arXiv},
eprintclass = {cs, eess},
doi = {10.48550/arXiv.2301.05841},
url = {http://arxiv.org/abs/2301.05841},
urldate = {2023-01-18},
abstract = {In this paper, we present a distributed optimal multiagent control scheme for quadrotor formation tracking under localization errors. Our control architecture is based on a leader-follower approach, where a single leader quadrotor tracks a desired trajectory while the followers maintain their relative positions in a triangular formation. We begin by modeling the quadrotors as particles in the YZ-plane evolving under dynamics with uncertain state information. Next, by formulating the formation tracking task as an optimization problem -- with a constraint-augmented Lagrangian subject to dynamic constraints -- we solve for the control law that leads to an optimal solution in the control and trajectory error cost-minimizing sense. Results from numerical simulations show that for the planar quadrotor model considered -- with uncertainty in sensor measurements modeled as Gaussian noise -- the resulting optimal control is able to drive each agent to achieve the desired global objective: leader trajectory tracking with formation maintenance. Finally, we evaluate the performance of the control law using the tracking and formation errors of the multiagent system.},
pubstate = {prepublished},
keywords = {Computer Science - Multiagent Systems,Electrical Engineering and Systems Science - Systems and Control},
}