Demo: ARpocalypse: Demonstrating Situational Awareness Challenges in Search-and-Rescue AR-HRC

Abstract

Augmented reality (AR) can improve human-robot collaboration (HRC) by spatially presenting robot state, intent, and task information, but such content can also introduce a security-relevant loss of situational awareness. We present ARpocalypse, an interactive demonstration in which participants collaborate with a robot in an emulated search-and-rescue scenario set amid an alien attack. The system uses large multimodal model (LMM)-guided, edge-assisted reasoning to identify candidate victim locations, visualizes robot navigation intent and sensing feedback in AR, and introduces an unexpected real-world threat that participants must detect and respond to. Through this experience, AR-pocalypse illustrates how AR-HRC visualizations can impose competing awareness demands, potentially reducing users' ability to perceive changes in the physical environment.