Firefighters' Perceptions on Collaboration and Interaction with Autonomous Drones: Results of a Field Trial

Title of the Paper

Firefighters’ Perceptions on Collaboration and Interaction with Autonomous Drones: Results of a Field Trial

Paper Information

• Subject Area: Human-Computer Interaction, Emergency Response Technology, Design and Evaluation of Highly Autonomous Drone Systems
• Keywords: Firefighting, Drones, Autonomous Systems, Human-Computer Interaction, Emergency Response, Task Collaboration, Information Overload, System Design, Field Trial, Sociotechnical

Research Background and Issues

• Identified Problems or Challenges:

  • The application of drones in firefighting and emergency response is increasing, but challenges remain in integrating them with firefighting teams and designing effective human-drone interaction methods.
  • Firefighters have expressed concerns about the visual and auditory interaction effects of drones, as well as issues related to information overload.
  • The use of drones may significantly alter existing firefighting processes and team structures, necessitating a reevaluation of these changes' impacts on training and decision-making.

• Importance of the Issues:

  • Firefighting environments are extremely hazardous, and optimizing safety measures can reduce casualties.
  • Drones have the potential to efficiently collect information, shorten response times, and perform dangerous tasks, making them potentially indispensable.
  • Researching the effectiveness of drones in real-world deployments and improving human-drone interaction design are critical to enhancing emergency response efficiency.

• Motivation and Related Work:

  • Investigating the sociotechnical collaboration requirements of drones in firefighting, such as trust issues and information transmission.
  • Limited studies have focused on the performance of drones in real-world operations; most prior research relies on surveys and simulated scenarios, which are insufficient to reflect the limitations and challenges of real-world operations.

Solution

• Proposed Solution:

  • Designed an interactive drone system based on light, sound, and graphical interfaces, featuring two types of drones (reconnaissance and search-and-rescue) to support firefighting tasks.
  • The system was tested through field trials (simulated building and forest fires) to achieve efficient collaboration and information support for firefighting missions.

• Innovative Aspects of the Solution:

  • Comprehensive evaluation of the interaction capabilities and impacts of drones by simulating fully autonomous drone system usage scenarios and conducting interviews with firefighters and observers.
  • Integration of multimodal interaction methods (visual light signals and audio broadcasts) while testing information flow processing functions, such as real-time video and message prioritization.

• Implementation Steps and Key Technologies:

  1. Designed two drone models: DJI Matrice 30 (supporting night vision and operation in adverse weather) and DJI Avata (for indoor search and rescue);
  2. Incrementally improved firefighting tasks by introducing rapid drone reconnaissance;
  3. Utilized real-time streaming, message notifications, and map interaction features for data transmission;
  4. Simulated fully autonomous drone operation during field trials using the "Wizard of Oz" method to collect firefighters’ reactions and feedback.

Research Findings

• Specific Findings:

  • Drone lighting helped firefighters quickly locate victims and improve nighttime operational efficiency, receiving widespread positive feedback.
  • Audio broadcasts faced challenges in high-noise environments due to unclear information, but appropriate optimization could enhance user experience.
  • Information overload remains a challenge; proposed solutions include introducing information storage tools (video playback, message prioritization) and delegating information to lower-level personnel.

• Advantages Over Existing Solutions:

  • Drones can arrive at fire scenes ahead of responders to collect environmental information, providing decision-making support for commanders.
  • The system’s multimodal interaction capabilities extend traditional information processing methods, reducing reliance on a single commander.

• Experimental and Evaluation Results:

  • Two field trials and 50 interviews revealed that drones accelerated task completion (saving 5-10 minutes) and significantly improved search-and-rescue efficiency.
  • Firefighters’ trust in autonomous drones requires further development, particularly regarding highly autonomous decision-making tasks, as they tend to prefer semi-autonomous human-drone collaboration solutions.

• Limitations and Future Directions:

  • Limitations: The study focused on building and forest fire scenarios, excluding other types of rescue missions; the sample was limited to Swiss volunteer firefighters, which may limit generalizability.
  • Challenges: Issues of information overload and the credibility of drone interactions with bystanders.
  • Future Directions: Expand research to diverse global firefighting contexts, explore the impact of multi-drone collaboration, and investigate drone applications in water rescue and complex tunnel missions.