[cosc-grad-students-list] RESEARCH PROPOSAL SEMINAR NOTICE-UNCERTAINTY AWARE PROBABILISTIC FORECASTING OF NON COOPERATIVE DYNAMIC OBSTACLES FOR SAFE AUTONOMOUS AERIAL NAVIGATION

[Eulenfeld, Menda]

RESEARCH PROPOSAL SEMINAR NOTICE

GEOSPATIAL COMPUTER SCIENCE PROGRAM

COLLEGE OF SCIENCE AND ENGINEERING

TEXAS A&M UNIVERSITY-CORPUS CHRISTI







TITLE: UNCERTAINTY‑AWARE PROBABILISTIC FORECASTING OF NON‑COOPERATIVE DYNAMIC OBSTACLES FOR SAFE AUTONOMOUS AERIAL NAVIGATION

SPEAKER: Syed Izzat Ullah

ADVISOR: Dr. Jose Baca

COMMITTEE: Dr. Pablo Rangel, Dr. Tianxing Chu, Dr. Carlos Rubio-Medrano, and Dr. Susan Elwood

DATE: June 24, 2025 (Tuesday)

TIME: 1:00 PM Central Time (US and Canada)

PLACE: CI-126

Zoom Link:  https://tamucc.zoom.us/j/94329044811?pwd=cxydgh9kmqARmq8nPaA0kg7LFqev2j.1<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Ftamucc.zoom.us%2Fj%2F94329044811%3Fpwd%3Dcxydgh9kmqARmq8nPaA0kg7LFqev2j.1&data=05%7C02%7Ccosc-grad-students-list%40listserv.tamucc.edu%7Cfce8f7d624cd4273b10608ddaa9283cf%7C34cbfaf167a64781a9ca514eb2550b66%7C0%7C0%7C638854268432303388%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=YN3WI5UGYCCmdVV02Cm9PAAtYz6AgenknZOR5kqRfYk%3D&reserved=0>

Meeting ID:  943 2904 4811

Meeting Passcode:  566848





Abstract



Unmanned Aerial Vehicles (UAVs) are increasingly deployed in complex, dynamic environments for applications such as urban air mobility, surveillance, emergency response, and package delivery. However, safe navigation in dynamic and uncertain environments featuring unpredictable obstacles and varying environmental conditions remains an unresolved challenge. Existing motion planning algorithms are constrained by significant limitations that undermine their effectiveness in real-world scenarios. First, current methods often rely on linear or constant-velocity models for obstacle motion prediction, which fail to capture the complex, nonlinear, and context-dependent behaviors of dynamic obstacles, leading to delayed or inadequate avoidance maneuvers and increased collision risks. Second, most trajectory planning frameworks lack modularity, treating obstacle prediction and planning as separate processes, which hinders proactive decision-making and limits the integration of advanced prediction models. Third, there is a critical gap in balancing the adaptability of learning-based methods, which excel at modeling complex obstacle behaviors, with robust safety mechanisms that account for uncertainties in perception (e.g., occlusions or noise in sensor readings) and environmental disturbances (e.g., wind gusts, turbulence, etc.), compromising reliability in safety-critical operations.

This research bridges these gaps through a unified probabilistic perception, data-driven forecasting, and verified trajectory optimization. First, we develop a real-time collision avoidance module that enhances state-of-the-art trajectory planners through short-term dynamic obstacle prediction using an Obstacle Filter. This transforms reactive systems into anticipation-aware architectures. Second, we develop and integrate deep learning-based prediction modules (e.g., Transformers, RNNs) to forecast multistep, probabilistic obstacle trajectories, enabling proactive avoidance of complex interactions. Third, we develop an uncertainty-aware module that incorporates perception uncertainties, and environmental disturbances into robust trajectory planning, ensuring reliability in challenging conditions like urban air mobility or search-and-rescue missions. Extensive simulations and hardware benchmarks will validate the framework’s improvements in safety, efficiency, and reliability. This research aims to bridge the critical gap in autonomous aerial systems, facilitating their reliable deployment in real-world applications where uncertainty is inherent.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.tamucc.edu/pipermail/cosc-grad-students-list/attachments/20250613/8bf6d18e/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/calendar
Size: 10236 bytes
Desc: not available
URL: <http://listserv.tamucc.edu/pipermail/cosc-grad-students-list/attachments/20250613/8bf6d18e/attachment-0001.ics>