2025 ARC Research Seminar - Winter Series

PI: Dr. Bradley Fahlman (Central Michigan U.)
(link to project)

Abstract: Due to the lightweight properties and desirable redox characteristics of lithium metal, lithium-ion batteries (LiBs) represent the most widely employed battery design for portable electronics and the transportation sector. However, in order to improve the range of EVs and decrease their overall weight, many other alternative battery chemistries have been in development. with metal-sulfur batteries an attractive choice. As one of the least expensive elements, sulfur is an attractive choice for battery cathodes with a theoretical specific capacity of 1675 mAh/g. This translates to a theoretical energy density for Li-S of 2510 Wh/kg, about 10-times larger than conventional Li-ion cells (250-300 Wh/kg). In addition to much larger energy densities than Li-ion batteries, Li-S batteries are less expensive ($90/kWh), relative to Li-ion ($140/kWh). Although much work has been expended on the development of metal-sulfur batteries, especially Li-S, the insulating behavior of sulfur and metal sulfides (e.g., Li₂S), the shuttling of polysulfides, as well as poor reversibility and slow kinetics of the metal-S conversion reactions, have limited their commercial applications. Herein, I will describe our work in improving the capacity and cyclability of Li-S batteries through the use of composite cathodes comprised of graphitic carbon nitride (g-C₃N₄), melt-infused with sulfur. Our preliminary work related to the synthesis and electrochemical testing of cathodes decorated with iron single atom catalysts (SACs) will also be discussed.

PI: Dr. Anja Mueller (Central Michigan U.)
(link to project)

Abstract: Different options for fuel cells are needed for autonomous ground vehicles in the battlefield. In this project, we are developing a proton exchange membrane (PEM) for high-temperature PEM fuel cells. The advantage of hydrogen fuel cells is that they are quiet and cooler than internal combustion engines, making them harder to detect. To optimize the proton transport in these new PEM materials, a continuous improvement cycle will be used that includes molecular dynamics simulation and coarse graining to predict optimized chemical structures, which will then synthesized and tested in the lab. The test results will then be used to further optimize the model. In this talk we will present our approach and initial results.

PI: Dr. Sang-Hwan Kim (U of M - Dearborn)
(link to project)

Abstract: This research explores optimal Human-Autonomy Teams (HAT) in dynamic dual-task environments, where AI supports secondary tasks. We designed a simulation featuring a primary skill-based motor task paired with a secondary rule-based decision-making task for target assessment. Initial experiments with human-human collaboration revealed natural task allocation and communication strategies. In the second year, we incorporated a pseudo-AI assistant, testing four task allocation strategies, two levels of AI decision-making transparency, and two communication modalities. Performance was assessed using metrics like speed, accuracy, workload, situation awareness, and trust. Results show that concise non-verbal information boosts performance when humans are more involved, whereas the take-over strategy - where control shifts between human and AI - harms situation awareness and trust due to mode transitions. These findings offer valuable insights for designing HAT systems in areas such as autonomous driving and complex mission execution.

PI: Dr. Alex Gorodetsky (U of M)
(link to project)

Abstract: In this talk, we will describe our recent development and deployment of incremental tensor decompositions for the purpose of enabling large scale data analysis. We begin by motivating the problem from the perspective of analyzing virtual gameplay to perform behavioral cloning of expert players. We will also highlight a motivating integration project with Oakland University whereby we seek to enable behavioral cloning for vehicles in a virtual game engine. Then we will describe the incremental algorithms that we have developed and compare their performance to existing state of the art - highlighting improvements in both speed, compression quality, and generalization. Finally, we will discuss some work in progress concerned with deploying these methods in the context of generative models.

PIs: Drs. Corina Sandu, Alba Yerro-Colom (Virginia Tech)
(link to project)

Abstract: As autonomy is integrated into combat vehicles, the need for high-fidelity virtual proving grounds becomes essential for accurate mobility prediction. This research focuses on modeling tire-soil interactions in saturated clays, which exhibit low shear strength, significant deformability, and complex failure mechanisms. In particular, the study examines the role of pore water in soil deformation and residual strength under shear loading. The time-scale effect of tire passes is also investigated to differentiate short-term and long-term multipass behavior. These goals are accomplished by, first, performing an experimental full-scale testing conducted at the Terramechanics Rig at Virginia Tech. Second, continuum-based modeling techniques are employed to simulate tire-soil interactions using an effective stress framework. A sensitivity analysis is conducted to identify key parameters driving the tire-soil system response. By benchmarking against new laboratory testing and state-of-the-art methods, this research proposes improved material modeling methodologies to capture large deformation behaviors more accurately. Future applications of these models will focus on multipass simulations, aiming to enhance the understanding of soil behavior beyond initial failure. The findings contribute to the advancement of predictive simulation tools, supporting the development of virtual proving grounds for autonomous vehicle mobility in challenging terrains.

PIs: Drs. Hiroyuki Sugiyama, Casey Harwood (U. of Iowa)
(link to project)

Abstract: In this talk, we will describe our recent development and deployment of incremental tensor decompositions for the purpose of enabling large scale data analysis. We begin by motivating the problem from the perspective of analyzing virtual gameplay to perform behavioral cloning of expert players. We will also highlight a motivating integration project with Oakland University whereby we seek to enable behavioral cloning for vehicles in a virtual game engine. Then we will describe the incremental algorithms that we have developed and compare their performance to existing state of the art --- highlighting improvements in both speed, compression quality, and generalization. Finally, we will discuss some work in progress concerned with deploying these methods in the context of generative models.