黑料吃瓜网

Monday, March 31, 2025.  Spanagel Hall 316,  12:00

The physics is against us: adventures in marine sensing systems

Erin M. Fischell, PhD (President and Founder, Acbotics Research LLC) 

Development of marine sensing systems is fundamentally limited and challenged by the ocean environment, from the physics of light attenuation to the vagaries of long-range acoustic propagation to the hazards presented by the sea ice. This presentation describes a series of projects that use one-way travel acoustic and seismic techniques to overcome these challenges while observing at a distance, for a diverse set of applications including AUV swarming, source of opportunity underwater track reconstruction, lineless lobster fishing, and ice crack localization. Physics may pose barriers to ocean sensing, but with modern electronics design and signal processing methods it can also be a tool to see a bit further in the ocean.

A Moore Inventor Fellow with more than 15 years marine robotics and underwater acoustics expertise, Dr. Fischell鈥檚 work spans propagation physics, arctic acoustics, autonomy, signal processing, instrumentation development, and robot perception topics. She received her PhD in Mechanical and Oceanographic Engineering from the MIT/WHOI Joint Program in 2015, and continued her work under a Battelle postdoctoral fellowship at MIT. From 2017-2021, Dr. Fischell was an Assistant Scientist at the Woods Hole Oceanographic Institution where she served as PI on AUV, autonomy and acoustics projects while teaching acoustics in the MIT-WHOI Joint Program. Dr. Fischell founded Acbotics Research with the objective of providing ubiquitous marine sensing systems to the scientific community; at Acbotics, she splits her time between subject matter consulting, product line development and testing, and continuing research in areas of acoustics, array processing, autonomy, and sensing.

Wednesday, March 5, 2025.  Spanagel Hall 316.   12:00

Optimizing Human-AI Teaming in High-Stakes Decision-Making through Empathy: A SONAR Operations Application

Madeline Facino,  University of Bath

Abstract 

The increasing integration of artificial intelligence (AI) in high-stakes decision-making, particularly within military contexts, has revealed significant barriers to trust and adoption. While AI often exceeds human capabilities in specific tasks, skepticism arises from unrealistic expectations of perfection, which hampers its acceptance in critical operations. This PhD project seeks to address these challenges in trust and adoption by investigating whether incorporating a model of empathy into Human-AI (HAI) interactions can foster more effective Human-AI Teaming (HAIT). By dynamically responding to human factors such as stress, cognitive overload, and fatigue, Empathetic Artificial Intelligence (EAI) has the potential to simulate human-like supportive behaviors - enhancing teamwork by facilitating the complementary strengths of both humans and AI. To test this hypothesis, we employ a Sound Navigation and Ranging (SONAR) simulation game, Tactical OpeRation Testbed for Underwater Gamified Analysis (TOrTUGA), designed to replicate high-stakes scenarios and integrate EAI behavior. Through an interdisciplinary approach that spans computer science, engineering, and social sciences, this research aims to determine whether EAI can reduce operator distress, enhance trust, and improve team performance in complex environments. Notably, this study is not focused on advancing AI technology itself, but on demonstrating how a model of empathy can transform HAIT by addressing the psychological and emotional needs of human operators.

Thursday, May 9, 2024.  Spanagel Hall, Room 316.  3-4pm 

Close-Range Remote Sensing Observations of Rocky Nearshore Hydrodynamics

Dr. Matthew Conlin, Oregon State University

Nearshore hydrodynamics on rough rocky shores are poorly understood relative to those on sandy shores. This is due primarily to a lack of observations, as in situ instrumentation is difficult to deploy in regions where semi- and fully-exposed rocks and roughness elements interact with shoaling and breaking waves. Close-range remote sensing is an attractive solution for the collection of nearshore hydrodynamic observations on rocky shores. In this presentation, I will provide a summary of hydrodynamic observations obtained on rocky shores around Monterey, and the insights they have provided, using unmanned aerial system-mounted lidar as well as mobile and shore-based video cameras. Observations include those of surface circulation in a rocky shore embayment, surge channel rip current dynamics, wave transformation and dissipation across a rocky surf zone, and shoreline wave reflection