Research Summary

(Selected Projects)

Jan 2010-, Principal Investigator, Raytheon Technologies Research Center(RTRC)/United Technologies Research Center(UTRC)

  • Dynamics and Controls on Hypergraphs (Co-PI, AFOSR Funded, Sep 2022-): Exploiting recent theoretical and computational advances in tensor algebra, hypergraph science and data driven systems and control theory to rigorously capture higher order interactions which underlie the interplay of structure, function, and dynamics inherent in complex adaptive systems. Applications are focused on biological problems.

  • Quantum Computing for Scientific Computing Applications (PI, RTRC IRAD, Aug 2022-): Developing new quantum algorithms for simulating high dimensional nonlinear ODE/PDEs on quantum platforms.

  • Learning Enabled Autonomy (CO-PI, RTRC IRAD, Jan 2021-): Applying state-of-the-art multi-agent reinforcement learning framework and algorithmic game theory for decision making in presence of uncertainty/adversaries in battlespace applications including sensor resource management and aerial combat. Developing distributed optimization and distributed estimation techniques for networked autonomous operations.

  • Insitu Monitoring and Control for Advanced/Additive Manufacturing (CO-PI, RTRC IRAD and Army/DOE funded, Jan 2018-): Developing machine learning/deep learning-based algorithms for automated defect detection for laser powder bed fusion process using variety of sensing modalities including photodiodes, high speed videos, and layer wise thermal and visual images. Applying advanced nonlinear estimation and reinforcement learning/control techniques for improving reliability/performance of advanced manufacturing processes including cold spray, thermoplastic composite joining.

  • Automating Inspection (CO-PI, IRAD, Aug 2016-): Developing ML/AI based algorithms for automating/aiding human inspection for a wide range of modalities including ultrasonic, visual and thermoacoustic imagery.

  • Multiway Dynamical Systems (Co-PI, AFOSR Funded, Mar 2018- Oct 2021): Developing a mathematical and computational framework for data driven modeling, analysis, and control of multiway dynamical systems as a new paradigm for characterizing complex systems using tensor-based approaches.

  • Learning for Dynamics and Controls (PI, UTRC IRAD, Jan 2015-Dec 2018): Interdisciplinary research combining techniques from Koopman operator theory, control theory and machine learning for data-driven aerospace applications including nonlinear estimation and control, prognostics and health management, and streaming analytics.

  • Supervisory Controller for Optimal Role Allocation for Cueing of Human Operators (PI, Army/ICB Funded, Aug 2014-Jun 2018): Developed novel sensor management and decision support tools to unburden the operator flying a rotorcraft, allowing him/her to conduct Unmanned Aerial System missions effectively while also conducting own-ship multitasking mission functions.

  • Vision-Based Autonomous Sensor-Tasking in Uncertain Adversarial Environments (PI, AFOSR Funded, Sep 2011- Aug 2014): Developed novel techniques for activity learning, detection and prediction in surveillance videos using a combination of methods from dynamical systems, control theory and statistics/machine learning.

  • Human Supervised Collaborative Autonomy (PI, UTRC IRAD, Jan 2011-Dec 2014): Developed new approaches for multi vehicle control and coordination for target search and tracking applications, multi agent task allocation under complex temporal constraints and uncertainties, and operator attention allocation optimization for supervisory control.

  • Sensor Tasking for Optimal Uncertainty Reduction (PI, UTRC IRAD, Jan-Dec 2010): Developed new techniques for anomaly detection, qualitative/quantitative information fusion, and sensor tasking optimization based on approximate dynamic programming methods. Demonstrated these techniques for real time camera network management for active multi target tracking in challenging indoor environments.

Aug 2007-Dec 2009, Individual Technical Contributor, United Technologies Research Center:

  • Decision Support Tool for Deep Energy Efficiency Retrofits (ESTCP (DoD) Funded): Developed a tool which can be used for rapid screening of economically viable energy efficiency retrofits at a portfolio level (e.g. large collection of buildings) to reduce energy consumption and minimize environmental impact.

  • DYNARUM (DARPA DSO Funded): Developed new techniques for scalable uncertainty quantification in large scale interconnected systems by combining ideas from spectral graph theory, waveform relaxation and traditional uncertainty quantification methods.

  • Integrated Buildings (UTRC Internally Funded): Research in the areas of building energy and security including sensor fusion and occupancy estimation, anomaly detection in people traffic, and reduced order modeling, estimation and control of building airflows.

Sep 2003-July 2007, Research Assistant, MIT:

  • Developed new dynamical system based methods for characterizing separation, mixing and transport in 3D unsteady fluid flows, and applied advanced complex analysis techniques for vortex dynamics in general multiply connected domains.

Aug 2001-Aug 2003, Research Assistant, PSU:

  • Research on discrete event control, anomaly detection, application of complex adaptive systems in design and analysis of multi-agent systems and supply chains, and dynamical system based analysis of reaction-diffusion PDEs.