See UD's plans for teaching, learning and research this fall with measures to promote safety and lessen the risk of COVID-19 spread. See UD case dashboard here.

Skip to main content

UD-UDRI Research Fellowship

UDRI Statements-of-Need

Ian Cannon - Reinforcement Learning Agent

Seeking Fellow Type: Junior

Description: UDRI is functioning as a trusted contractor for a variety DoD projects in a variety of topic areas, including autonomy and autonomous systems, human machine teaming, artificial intelligence, data exploitation, machine and deep learning, target detection and tracking, and other related capabilities. To stay competitive in the field, UDRI proposes a research and development effort to advance the state-of-the-art in reinforcement learning (RL). Reinforcement learning is a pillar of machine learning in which a software agent learns to interact with its environment. A research fellow is required to assist in the verification of an RL environment and development of a functioning agent and to provide expert inputs for future work on the environment and agent. UDRI would also like to tap into the considerable expertise of University of Dayton faculty to generate unique and innovative ideas in reinforcement learning and related research areas.

Proposal: This project offers the School of Engineering faculty an opportunity to generate novel, innovative ideas in the emerging field of reinforcement learning and related research areas. It also offers a collaborative opportunity to generate white papers, research papers and submit proposals to a variety of DoD topics/RFIs/RFPs based on those ideas. Successful proposals provide important contract vehicles that may be utilized by UDRI and UD's School of Engineering and College of Arts & Sciences.

Michael Hanchak - Thermal Modeling of High-Speed Vehicle Power Extraction

Seeking Fellow Type: Junior

Description: The UDRI Modeling & Simulation group has a need for a 2020 summer faculty member to perform numerical modeling of power extraction components for potential use in a high-speed vehicle. This work would be performed at WPAFB in direct support of AFRL/RQQM.

The fellow should possess a Ph.D. in the thermal sciences and have experience creating numerical models in either a popular programming language or commercial analysis software.

The fellow should have good communication skills and a positive, enthusiastic attitude. Furthermore, the candidate must be a U.S. citizen.

I would prefer that the fellow work at least two days per week during the 2020 summer session. This work will take place in Building 18, Area B, WPAFB. Some in-processing may be required. 

Proposal: This fellowship would be an excellent opportunity to introduce a new faculty to the Air Force Research Laboratory. UDRI has a rich history of collaboration with AFRL. I feel this relationship should be explored, whenever possible, to provide future long-term funding options for UD faculty.

This fellowship would introduce a newer faculty to the problems facing AFRL. Furthermore, this would help form some critical relationships among AFRL, UDRI and UD collaborators. UDRI has a distinct advantage over its competitors precisely because of programs like this that can provide academic assistance at no cost to AFRL. Our group at UDRI can provide some additional funds if a computer workstation is needed, for example.

As high-speed vehicles generate an enormous amount of heat on their journey through the atmosphere, AFRL is looking for innovative ways of extracting some of that thermal energy into electrical energy. The fellow can leverage existing research as well as providing innovative solutions and research. It is likely that this effort could result in a publication, provided that it passes the AF public release process.

Zhenhua Jiang - Foundations and Applications of Trustworthy AI in Physical Systems

Seeking Fellow Type: Junior & Senior

Description: The long-term goal of this effort is to investigate the foundations and applications of trustworthy AI in physical systems. The physical systems of interest include mobile robots, drones, unmanned aerial, ground or underwater vehicles, aircraft electrical power systems, microgrids, facility or industrial energy systems, infrastructure systems, experimental systems for material discovery and testing, additive manufacturing systems, and command and control systems.

One example area is to research, develop, test and evaluate a real-time, adaptable autonomy kernel on an open-architecture, FPGA-based edge computing platform, to enable a variety of intelligent autonomous systems. These systems may operate in a very dynamic and uncertain environment. A probabilistic computational framework is proposed to address the uncertainty quantification and operation optimization in an autonomous manner. Artificial intelligence (AI) methods, such as long short-term memory (LSTM) neural networks, generative adversarial networks (GAN), spiking neural networks (SNN), convolutional neural networks (CNN), etc., can be used to characterize and predict the temporal and spatial dynamics of the autonomous systems, optimize the real-time operation considering constraints and support or automate the online decision-making.

To achieve the long-term goal, technical skills needed include machine learning, knowledge representation, reasoning, inference, probability theory, operations research, cybersecurity, real-time computing, adaptive control, communication networks, as well as infrastructure systems, experimental systems for material discovery and testing, additive manufacturing systems, and command and control systems (e.g., tracking, targeting, engagement, decision-making, etc.).

Proposal: Several medium-to-large white papers or full proposals (in multi-million dollars each) are planned for the next year in multiple application areas: 

  • NSF – AI Institute: Foundations and Applications of Trustworthy AI in Physical Systems ($20M)
  • ARPA-E – Machine Learning-Enabled Energy Product Development
  • AFRL/RI - Mastering Complexity in Multi-Domain Command and Control
  • DARPA - Artificial Intelligence Exploration
  • ESTCP – An Intelligent Multi-Microgrid Controller for Military Installations
  • AFRL/RQQ - Integrated Control for Aircraft Vehicle Energy System Optimization
  • ONR or NAVSEA - Real-Time Hardware-Accelerated Adaptable Autonomy Kernel for Naval Use

Paul Kladitis - Textile Batteries for Multifunctional Structures and Materials

Description: A multifunctional structure is a mechanically robust structure that is efficient with respect to size and weight but has additional functionality built inside. One of the most elusive functions to incorporate seamlessly into a structure is a power supply. The resultant structure must also be safe and ideally bio-compatible with humans for wearable or health related applications. I am looking for expertise in cloth-based batteries and biological applications to assist me with designing integrated power using my textile-based structure manufacturing equipment.

Proposal: From this effort we hope to use the knowledge gained for proposals. The information obtained during this period will also be enough for technical publications.

Victoria Kramb - Machine Learning Applications for Complex Materials Characterization

Seeking Fellow Type: Junior & Senior

Description: The NDE Engineering group has developed innovative methods for nondestructive detection of flaws and characterization of material properties using various sensing technologies. Thus far, transition of these sensing approaches to the end user has been limited to discrete flaw detection within somewhat isotropic materials. Advanced materials, such as those produced using additive manufacturing, multilayer composite structures and hybrid materials exhibit large variations in microstructure, macroscopic features and material properties, which result in an inability to distinguish between flaws and normal material variability during inspections. Innovative methods of data processing are needed to perform comparative studies between characterization data on acceptable, as-processed material and that which contains known flaws so that robust flaw detection criteria could be developed. Once metrics have been identified that can identify regions with high probability of flaws, additional studies could be performed to determine flaw characteristics such as physical dimensions and depth in the component.

Proposal: This research proposal seeks to establish a collaboration between the NDE Engineering group and a CAS or SoE faculty member with experience in machine learning to develop a data analysis framework for evaluation of NDE sensor data. Once developed, this framework will provide for optimized analysis of the large quantity of NDE data necessary for application of machine learning processes. The proposed collaboration would result in processes applicable to sensor data for applications in DoD for safety and sustainability of aircraft and in manufacturing for improved part and material quality. Further, integration of the machine learning framework into existing data acquisition systems would allow for development of a self-teaching process that would lead to optimized flaw detection processes and development of new sensing systems for specific applications. DoD, NASA and NSF researchers have already identified machine learning processes as the next generation approach to big data analytics. The partnership established with this collaboration will allow UDRI to further advance our position within the NDE research and machine learning communities to secure funding for follow-on efforts through SBIR and AFOSR funding.

Due to ITAR restrictions on the research data being studied, participation is restricted to U.S. citizens or Green Card holders.

Victoria Kramb - Mixed Reality Integration for Enhanced Operator Safety and Reliability

Seeking Fellow Type: Junior & Senior

Description: The high reliability and safety of aviation travel is made possible through structured and repeatable inspections performed by maintenance personnel on aircraft structures and engines. As the inspected components and instrumentation become more complex, the need for in-depth training programs becomes essential in order to maintain the high level of aircraft reliability expected by air travelers. Since many aircraft structures cannot be reproduced within a classroom environment, mixed reality provides the ideal scenario for hands-on training utilizing actual instrumentation integrated with virtual structures. Capability has already been developed for integration of mixed reality structures and simulated sensor data. The next generation mixed reality hardware will make hands-on training modules possible with the integration of hand motion tracking with actual sensor data and positioning on the simulated aircraft component. 

Proposal: The NDE Engineering group demonstrated capability developed under the previous faculty collaboration project to DoD and internal customers for NDE training applications. The initial development included integration of 3D models that could be displayed in the field of view for the user. Further development is needed to integrate the instrument outputs so that the user can view inspection data as it is mapped onto the simulated part. Once this capability has been demonstrated, opportunities for development of NDE training modules are anticipated through current DoD and NDE industrial customers. Investments for the Hololens II equipment have already been made through NDE Engineering internal funding but further collaboration with the CAS is needed to fully integrate and demonstrate the proposed capability.

Due to the DoD connections with potential NDE studies, a U.S. citizen or Green Card holder is required for this position.


School of Engineering Associate Deans

Kettering Laboratories
300 College Park
Dayton, Ohio 45469 - 0254