Author Information

Is this project an undergraduate, graduate, or faculty project?

Undergraduate

Project Type

individual

Campus

Daytona Beach

Authors' Class Standing

Taylor Hostetter, Senior

Lead Presenter's Name

Taylor Hostetter

Lead Presenter's College

DB College of Engineering

Faculty Mentor Name

Dr. Kranthi Kumar Deveerasetty

Abstract

Lightweight UAV-to-UAV Detection and Tracking for Advanced Air Mobility Applications addresses the significant challenge of reliable UAV-to-UAV detection on resource-constrained platforms, particularly within Advanced Air Mobility (AAM) environments where dense, low-altitude airspace requires robust detect-and-avoid capabilities. This work presents the development and experimental evaluation of a lightweight detection and tracking framework for autonomous detect-and-avoid applications. The approach is designed to support real-time onboard operation in multi-vehicle environments characteristic of emerging AAM systems. The proposed framework integrates optical and LiDAR sensing with a low-complexity machine learning decision-support layer that reduces false detections without replacing the underlying control-oriented detection pipeline. This design preserves deterministic behavior and computational efficiency, both of which are critical for scalable deployment in AAM platforms. Experimental validation includes both stationary and moving-frame testing under varying motion, background clutter, and environmental conditions, using quantitative metrics to evaluate detection accuracy and robustness. Preliminary results show that a preprocessing dehazing algorithm improves detection performance in low-visibility conditions by increasing the consistency of target identification. A YOLOv8-based detector, integrated alongside the primary detection pipeline, improves recall, particularly for stationary UAVs that are difficult to detect using motion-based methods alone. Most notably, a post-processing LSTM-based tracker significantly reduces false track persistence, resulting in the largest improvement in overall precision. This work demonstrates that machine learning can be selectively integrated to enhance system robustness while preserving the computational efficiency and deterministic behavior required for real-time control. These characteristics are essential for enabling reliable multi-UAV sensing, closed-loop collision avoidance, and scalable airspace integration in Advanced Air Mobility systems.

Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?

No

Share

COinS
 

Lightweight UAV-to-UAV Detection and Tracking for Advanced Air Mobility Applications

Lightweight UAV-to-UAV Detection and Tracking for Advanced Air Mobility Applications addresses the significant challenge of reliable UAV-to-UAV detection on resource-constrained platforms, particularly within Advanced Air Mobility (AAM) environments where dense, low-altitude airspace requires robust detect-and-avoid capabilities. This work presents the development and experimental evaluation of a lightweight detection and tracking framework for autonomous detect-and-avoid applications. The approach is designed to support real-time onboard operation in multi-vehicle environments characteristic of emerging AAM systems. The proposed framework integrates optical and LiDAR sensing with a low-complexity machine learning decision-support layer that reduces false detections without replacing the underlying control-oriented detection pipeline. This design preserves deterministic behavior and computational efficiency, both of which are critical for scalable deployment in AAM platforms. Experimental validation includes both stationary and moving-frame testing under varying motion, background clutter, and environmental conditions, using quantitative metrics to evaluate detection accuracy and robustness. Preliminary results show that a preprocessing dehazing algorithm improves detection performance in low-visibility conditions by increasing the consistency of target identification. A YOLOv8-based detector, integrated alongside the primary detection pipeline, improves recall, particularly for stationary UAVs that are difficult to detect using motion-based methods alone. Most notably, a post-processing LSTM-based tracker significantly reduces false track persistence, resulting in the largest improvement in overall precision. This work demonstrates that machine learning can be selectively integrated to enhance system robustness while preserving the computational efficiency and deterministic behavior required for real-time control. These characteristics are essential for enabling reliable multi-UAV sensing, closed-loop collision avoidance, and scalable airspace integration in Advanced Air Mobility systems.

 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.