Applying Ground Robot Fleets in Urban Search: Understanding Professionals' Operational Challenges and Design Opportunities

Urban searches demand rapid, defensible decisions and sustained physical effort under high cognitive and situational load. Incident commanders must plan, coordinate, and document time-critical operations, while field searchers execute evolving tasks in uncertain environments. With recent advances in technology, ground-robot fleets paired with computer-vision-based situational awareness and LLM-powered interfaces offer the potential to ease these operational burdens. However, no dedicated studies have examined how public safety professionals perceive such technologies or envision their integration into existing practices, risking building technically sophisticated yet impractical solutions. To address this gap, we conducted focus-group sessions with eight police officers across five local departments in Virginia. Our findings show that ground robots could reduce professionals’ reliance on paper references, mental calculations, and ad-hoc coordination, alleviating cognitive and physical strain in four key challenge areas: (1) partitioning the workforce across multiple search hypotheses, (2) retaining group awareness and situational awareness, (3) building route planning that fits the lost-person profile, and (4) managing cognitive and physical fatigue under uncertainty. We further identify four design opportunities and requirements for future ground-robot fleet integration in public-safety operations: (1) scalable multi-robot planning and control interfaces, (2) agency-specific route optimization, (3) real-time replanning informed by debrief updates, and (4) vision-assisted cueing that preserves operational trust while reducing cognitive workload. We conclude with design implications for deployable, accountable, and human-centered urban-search support systems

💡 Research Summary

This paper investigates how ground‑robot fleets could be integrated into urban lost‑person search operations, focusing on the perspectives of frontline public‑safety professionals. Recognizing that such searches demand rapid, defensible decisions under high cognitive and physical load, the authors note that current practices remain largely manual, relying on paper references, mental calculations, and ad‑hoc coordination. To bridge the gap between advanced robotic capabilities and operational realities, the researchers conducted focus‑group sessions with eight police officers from five Virginia law‑enforcement agencies.

Through qualitative analysis, four high‑level challenge areas emerged: (1) partitioning the workforce across multiple search hypotheses, (2) maintaining shared situational awareness among dispersed teams, (3) constructing route plans that align with the lost‑person’s profile, and (4) managing cognitive and physical fatigue amid uncertainty. Participants repeatedly emphasized that robots could alleviate reliance on paper manuals and reduce the mental bookkeeping required to keep teams synchronized.

From these insights, the authors derive four design opportunities for future human‑centered robot‑fleet systems:

1. Scalable Multi‑Robot Planning and Control Interfaces – A unified dashboard that visualizes the status, trajectories, and sensor feeds of many robots, allowing commanders to issue batch commands and monitor progress without overwhelming their attention.

2. Agency‑Specific Route Optimization – Integration of locally relevant data (road restrictions, historical disappearance patterns, terrain constraints) so that the fleet can automatically generate routes tailored to the demographic and behavioral profile of the missing individual.

3. Real‑Time Replanning Informed by Debrief Updates – Mechanisms for ingesting new field information—witness statements, weather changes, or emerging clues—and instantly updating robot assignments and search zones, thereby keeping the operation adaptive.

4. Vision‑Assisted Cueing with Trust‑Preserving Explanations – Computer‑vision modules that flag persons of interest and provide confidence scores, coupled with large‑language‑model (LLM) generated textual rationales that help operators verify and, if needed, override robot suggestions.

The paper stresses that any deployed system must satisfy accountability and trust requirements: robot decisions need transparent logs, explainable outputs, and the ability for human operators to intervene. Moreover, the design must accommodate agencies of varying size and resources, suggesting modular, cloud‑enabled architectures that can scale from small precincts to larger departments.

In sum, the study contributes a rare, practitioner‑grounded view of how ground‑robot fleets could reduce cognitive load, improve coordination, and enhance the defensibility of urban search missions. By mapping concrete operational pain points to concrete HCI, robotics, and AI design requirements, the authors provide a roadmap for researchers aiming to create deployable, accountable, and human‑centric search‑support technologies.