Vision

Field-aligned XR guidance for uncertain public-safety terrain.

A² Perceptions is translating shared-perception research into a SAR-first platform that turns fragmented mission data into spatial guidance, confidence-aware cues, and team coordination support.

Research Insights Technical Direction
01

Field XR needs privacy and trust designed into the platform from the beginning.

BystandAR demonstrated a practical AR privacy-protection approach that uses eye gaze, voice, and spatial awareness to distinguish interaction subjects from bystanders. In a 16-participant study, the system protected 98.14% of bystanders while preserving 96.27% subject availability and maintaining 52.6 FPS with on-device processing.

Product implication

A² should treat privacy, data minimization, and on-device processing as core design principles for responder XR—not as later compliance add-ons.

02

Shared perception can reduce the time needed to locate critical information.

Ajna showed that shared spatial AR references can help users distribute detections and spatial context without relying on QR codes, extra edge hardware, or cloud services. In a simulated SAR study, participants using Ajna located victims 15% faster, 89% reported improved situational awareness, and 13 of 15 rated the system acceptable for usability.

Product implication

A² should focus less on novelty overlays and more on practical shared spatial references: teammate locations, clue markers, last-seen points, hazards, and routes that can be acted on quickly.

03

Remote sensors can help, but only if they reduce—not add—sensemaking burden.

KHAIT showed how AI-enabled canine perspective data can help handlers interpret where a survivor may be and navigate more efficiently, with an average 22% time-efficiency improvement in tested trials. The results also varied by scenario, canine, and user interaction, reinforcing that field systems must be designed for uncertainty and operational variation.

Product implication

A² should treat drones, K9 cameras, trail cameras, and teammate feeds as mission inputs that need summarization, prioritization, and confidence cues—not just more video windows.

04

Distributed AI can speed consensus, but trust must be made visible.

The distributed-AI SAR study used 24 subject-matter experts and found that AI-assisted teams reached decision consensus faster than controls. The same study emphasized that AI support introduces new HCI challenges: trust cues, conflicting detections, attention management, and the need to turn detections into shared situational awareness without overwhelming the team.

Product implication

A² should communicate confidence, uncertainty, and corroboration directly in the workflow so human judgment remains central during ambiguous missions.

Technical direction

Research evidence points toward a focused platform architecture.

A² is moving toward a SAR-first system that turns uncertain field data into interpretable spatial guidance. The direction is based on what the research repeatedly surfaced: location must be useful even when imperfect, mission data must become shared context, and AI cues must communicate uncertainty rather than imply certainty.

01

Hybrid outdoor localization

Blend GPS/GNSS, visual-inertial tracking, spatial anchoring, and map-based references so guidance remains useful when location estimates drift.

02

Mission-data interoperability

Transform routes, search assignments, hazards, clues, teammate locations, and boundaries into XR-ready spatial cues.

03

Uncertainty-aware cueing

Show whether cues are precise, approximate, degraded, or corroborated so responders can interpret guidance without over-trusting automation.

Core platform question

Can uncertain mission data become field-aligned cues responders can interpret quickly?

The answer will shape whether A² can mature from research foundation to repeatable public-safety platform.

Research to commercialization timeline

How the research foundation grows into the A² platform vision.

The publication timeline shows how the team’s research connects to the commercial direction: privacy-preserving AR, shared spatial perception, human/AI teaming, distributed SAR collaboration, and the current path toward field-aligned XR guidance.

2023

Privacy-first AR

BystandAR

Demonstrated real-time, on-device protection of bystander camera and depth data using gaze, voice, and spatial awareness. This supports A²’s need to build field XR around privacy, trust, and edge processing.

Corbett et al., MobiSys 2023
2025

Shared spatial perception

Ajna

Showed that wearable shared perception can improve team sensemaking, reduce search time, and increase situational awareness in emergency-response scenarios.

Wilchek et al., ACM TiiS 2025
2025

Human/AI/K9 teaming

KHAIT

Explored AI-supported canine-team sensemaking, showing how remote sensing and AI cues can help handlers interpret survivor location evidence more efficiently.

Wilchek et al., IUI 2025
2026

Distributed AI collaboration

Distributed AI for SAR

Examined how distributed AI affects trust, collaboration, and consensus for SAR teams, reinforcing the need for confidence-aware cues and human-centered AI coordination.

Wilchek et al., CHI 2026
Now

Lab to launch

A² Perceptions commercialization path

Responder-centered discovery and early commercialization planning translate the research lineage into a SAR-first platform direction focused on hybrid localization, mission-data interoperability, and uncertainty-aware XR cueing.

View the technical direction

Next step

Help pressure-test the launch path.

We welcome conversations with SAR teams, public-safety professionals, emergency managers, researchers, and technology partners who can help refine the first capability set and future pilots.

Connect With Us