NODEO: A Neural Ordinary Differential Equation Based Optimization Framework for Deformable Image Registration

Y. Wu, T. Z. Jiahao, J. Wang, P. A. Yushkevich, M. A. Hsieh, J. C. Gee

Abstract: Deformable image registration (DIR), aiming to find spatial correspondence between images, is one of the most critical problems in the domain of medical image analysis. In this paper, we present a novel, generic, and accurate diffeomorphic image registration framework that utilizes neural ordinary differential equations (NODEs). We model each voxel as a moving particle and consider the set of all voxels in a 3D image as a high-dimensional dynamical system whose trajectory determines the targeted deformation field. Our method leverages deep neural networks for their expressive power in modeling dynamical systems, and simultaneously optimizes for a dynamical system between the image pairs and the corresponding transformation. Our formulation allows various constraints to be imposed along the transformation to maintain desired regularities. Our experiment results show that our method outperforms the benchmarks under various metrics. Additionally, we demonstrate the feasibility to expand our framework to register multiple image sets using a unified form of transformation,which could possibly serve a wider range of applications.

Status: Accepted to CVPR 2022.

Preprint available on arXiv:

Topology Control of a Periodic Time-varying Communication Network with Stochastic Temporal Links

L. Shen, X. Yu, and M. A. Hsieh

Abstract: Mobile agents can form communication networks with links that emerge and disappear over time. Information transmission on such networks must pass through a sequence of links that are activated in a certain chronological order. Uncertainties in a link’s activation or deactivation risk violating the chronological order of the formation of links in paths for information transmission. The reasons for these uncertainties can be varied, but their presence prevents the application of existing topology control tools for mitigation. We propose an approach to leverage the pre-stored knowledge of the scheduled links’ events and create time-respecting subpaths accordingly. By estimating the impact of the stochastic timing on each subpath and by measuring how influential the subpaths are, we are able to measure the impact of uncertainties on any link event’s timing on the whole network’s information transmission performance. Reducing the uncertainty in the timing of those links with the highest impact can largely improve the network’s performance. Our method is tested and validated with simulation results.

Status: Accepted to ACC 2022.

Preprint available on arXiv: To come.

Flow-Based Control of Marine Robots in Gyre-Like Environments

G. Knizhnik, P. Li, X. Yu, and M. A. Hsieh

Abstract: We present a flow-based control strategy that enables resource-constrained marine robots to patrol gyre-like flow environments on an orbital trajectory with a periodicity in a given range. The controller does not require a detailed model of the flow field and relies only on the robot’s location relative to the center of the gyre. Instead of precisely tracking a pre-defined trajectory, the robots are tasked to stay in between two bounding trajectories with known periodicity. Furthermore, the proposed strategy leverages the surrounding flow field to minimize control effort. We prove that the proposed strategy enables robots to cycle in the flow satisfying the desired periodicity requirements. Our method is tested and validated both in simulation and in experiments using a low-cost, underactuated, surface swimming robot, i.e. the Modboat.

Status: Accepted to ICRA 2022.

Preprint available on arXiv:

Learning to Swarm with Knowledge-Based Neural Ordinary Differential Equations

T. Z. Jiahao, L. Pan, and M. A. Hsieh

Abstract: Understanding decentralized dynamics from collective behaviors in swarms is crucial for informing robot controller designs in artificial swarms and multiagent robotic systems. However, the complexity in agent-to-agent interactions and the decentralized nature of most swarms pose a significant challenge to the extraction of single-robot control laws from global behavior. In this work, we consider the important task of learning decentralized single-robot controllers based solely on the state observations of a swarm’s trajectory. We present a general framework by adopting knowledge-based neural ordinary differential equations (KNODE) — a hybrid machine learning method capable of combining artificial neural networks with known agent dynamics. Our approach distinguishes itself from most prior works in that we do not require action data for learning. We apply our framework to two different flocking swarms in 2D and 3D respectively, and demonstrate efficient training by leveraging the graphical structure of the swarms’ information network. We further show that the learnt single-robot controllers can not only reproduce flocking behavior in the original swarm but also scale to swarms with more robots.

Status: Accepted to ICRA 2022.

Preprint available on arXiv:

KNODE-MPC: A Knowledge-Based Data-Driven Predictive Control Framework for Aerial Robots

K. Y. Chee, T. Z. Jiahao and M. A. Hsieh

Abstract: In this letter, we consider the problem of deriving and incorporating accurate dynamic models for model predictive control (MPC) with an application to quadrotor control. MPC relies on precise dynamic models to achieve the desired closed-loop performance. However, the presence of uncertainties in complex systems and the environments they operate in poses a challenge in obtaining sufficiently accurate representations of the system dynamics. In this letter, we make use of a deep learning tool, knowledge-based neural ordinary differential equations (KNODE), to augment a model obtained from first principles. The resulting hybrid model encompasses both a nominal first-principle model and a neural network learnt from simulated or real-world experimental data. Using a quadrotor, we benchmark our hybrid model against a state-of-the-art Gaussian Process (GP) model and show that the hybrid model provides more accurate predictions of the quadrotor dynamics and is able to generalize beyond the training data. To improve closed-loop performance, the hybrid model is integrated into a novel MPC framework, known as KNODE-MPC. Results show that the integrated framework achieves 60.2% improvement in simulations and more than 21% in physical experiments, in terms of trajectory tracking performance.

Status: Published in IEEE RA-L.

Preprint available on arXiv:

Learning and Leveraging Features in Flow-Like Environments to Improve Situational Awareness

T. Salam, V. Edwards, and M. A. Hsieh

Abstract: This letter studies how global dynamics and knowledge of high-level features can inform decision-making for robots in flow-like environments. Specifically, we investigate how coherent sets, an environmental feature found in these environments, inform robot awareness within these scenarios. The proposed approach is an online environmental feature generator which can be used for robot reasoning. We compute coherent sets online with techniques from machine learning and design frameworks for robot behavior that leverage coherent set features. We demonstrate the effectiveness of online methods over offline methods. Notably, we apply these online methods for robot monitoring of pedestrian behaviors and robot navigation through water. Environmental features such as coherent sets provide rich context to robots for smarter, more efficient behavior.

Status: Published in IEEE RA-L.

Preprint available on arXiv:

Come see our work at ICRA 2022!

Come and see our presentations at ICRA 2022 next week (May 23-27)!

  • Tahiya Salam, Sandeep Manjanna, Ani Hsieh, and Gregory Dudek from McGill University will be hosting the Robotics for Climate Change Workshop on Monday in Room 108B. See here for more information!
  • Kong Yao Chee will be presenting our KNODE-MPC work in the Motion Control Session (TuA16) on Tuesday morning.  His talk is TuA16.07 and starts at 10:45 in Room 122B.
  • Gedaliah Knizhnik will be presenting his work with Mark Yim on control for docking Modboats in the Marine Robotics/Localization Session (TuB13) on Tuesday afternoon. His talk is TuB13.01 and starts at 15:30 in Room 116.
  • Peihan Li will be presenting our work on controlling Modboats in ocean-like environments in the Marine Robotics/Localization Session (TuB13) on Tuesday afternoon. His talk is TuB13.04 and starts at 15:45 in Room 116.
  • Tahiya Salam will be presenting our work on learning transfer operators using kernel methods in the Industrial and Environmental Robotics and Monitoring Session (WeA13) on Wednesday morning. Her talk is WeA13.01 and starts at 10:00 in Room 116.
  • Tom Zhang will be presenting our work with learning swarming controllers using KNODE in the Multi-Robot and Swarm Robotics II Session (WeB12) on Wednesday afternoon. His talk is WeB12.09 and starts at 16:20 in Room 118C.
  • Come see Jasleen Dhanoa and her MEAM 520 Fall 2022 cohorts demoing their MEAM 520 final project at the GRASP Exhibit Booth in the Exhibit Hall Tue-Thu!