Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Real-Time Assessment of Bystander Situation Awareness in Drone-Assisted First Aid

Shen Chang, Renran Tian, Nicole Adams, Nan Kong

TL;DR

This paper tackles the challenge of real-time situational awareness (SA) for bystanders in drone-assisted first aid during opioid overdose emergencies. It introduces the Drone-Assisted Naloxone Delivery Simulation Dataset (DANDSD) and a two-stage SA assessment framework that fuses perception, disparity, graph embeddings, and transformer-based imitation learning to predict SA in real time. The approach achieves state-of-the-art performance in binary and ternary SA prediction and significantly improves temporal segmentation accuracy over the FINCH baseline, demonstrating robust, interpretable SA dynamics aligned with human cognition. The results hold promise for adaptive drone guidance that enhances bystander effectiveness and emergency outcomes in time-critical OHME scenarios.

Abstract

Rapid naloxone delivery via drones offers a promising solution for responding to opioid overdose emergencies (OOEs), by extending lifesaving interventions to medically untrained bystanders before emergency medical services (EMS) arrive. Recognizing the critical role of bystander situational awareness (SA) in human-autonomy teaming (HAT), we address a key research gap in real-time SA assessment by introducing the Drone-Assisted Naloxone Delivery Simulation Dataset (DANDSD). This pioneering dataset captures HAT during simulated OOEs, where college students without medical training act as bystanders tasked with administering intranasal naloxone to a mock overdose victim. Leveraging this dataset, we propose a video-based real-time SA assessment framework that utilizes graph embeddings and transformer models to assess bystander SA in real time. Our approach integrates visual perception and comprehension cues--such as geometric, kinematic, and interaction graph features--and achieves high-performance SA prediction. It also demonstrates strong temporal segmentation accuracy, outperforming the FINCH baseline by 9% in Mean over Frames (MoF) and 5% in Intersection over Union (IoU). This work supports the development of adaptive drone systems capable of guiding bystanders effectively, ultimately improving emergency response outcomes and saving lives.

Real-Time Assessment of Bystander Situation Awareness in Drone-Assisted First Aid

TL;DR

This paper tackles the challenge of real-time situational awareness (SA) for bystanders in drone-assisted first aid during opioid overdose emergencies. It introduces the Drone-Assisted Naloxone Delivery Simulation Dataset (DANDSD) and a two-stage SA assessment framework that fuses perception, disparity, graph embeddings, and transformer-based imitation learning to predict SA in real time. The approach achieves state-of-the-art performance in binary and ternary SA prediction and significantly improves temporal segmentation accuracy over the FINCH baseline, demonstrating robust, interpretable SA dynamics aligned with human cognition. The results hold promise for adaptive drone guidance that enhances bystander effectiveness and emergency outcomes in time-critical OHME scenarios.

Abstract

Rapid naloxone delivery via drones offers a promising solution for responding to opioid overdose emergencies (OOEs), by extending lifesaving interventions to medically untrained bystanders before emergency medical services (EMS) arrive. Recognizing the critical role of bystander situational awareness (SA) in human-autonomy teaming (HAT), we address a key research gap in real-time SA assessment by introducing the Drone-Assisted Naloxone Delivery Simulation Dataset (DANDSD). This pioneering dataset captures HAT during simulated OOEs, where college students without medical training act as bystanders tasked with administering intranasal naloxone to a mock overdose victim. Leveraging this dataset, we propose a video-based real-time SA assessment framework that utilizes graph embeddings and transformer models to assess bystander SA in real time. Our approach integrates visual perception and comprehension cues--such as geometric, kinematic, and interaction graph features--and achieves high-performance SA prediction. It also demonstrates strong temporal segmentation accuracy, outperforming the FINCH baseline by 9% in Mean over Frames (MoF) and 5% in Intersection over Union (IoU). This work supports the development of adaptive drone systems capable of guiding bystanders effectively, ultimately improving emergency response outcomes and saving lives.

Paper Structure

This paper contains 28 sections, 1 equation, 4 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Traditional Situational Awareness (SA) Studies
  4. AI-Assisted SA Assessment
  5. Temporal Segmentation
  6. Methodology
  7. Time Interval Annotations
  8. SA Rating Annotations
  9. SA Prediction
  10. Overall Framework via Compositional Learning
  11. Perception Module
  12. Comprehension Module
  13. Disparity Estimation
  14. Graph Embedding
  15. Transformer-based Imitation Learning
  16. ...and 13 more sections

Figures (4)

  • Figure 1: Observer-rated SA changes during bystander-drone interaction using single-scale measurement.
  • Figure 2: Configuration for the overall perception and comprehension modules via compositional learning.
  • Figure 3: SA prediction performance on the testing sample. The solid line represents the SA prediction output generated by the well-trained Ternary classification model using all concatenated features as input. The dashed line depicts the output after smoothing with a 13-frame Gaussian filter, corresponding to the human reaction time of 0.25 seconds.
  • Figure 4: Qualitative Temporal Segmentation Results across all methods. "GT" denotes expert observations of event boundaries based on changes in bystander's SA.