Evidence of Absence: Cross-Modal Abductive Risk Perception to Sustain World Models When Vision Fails

2026-08-14

AI Summary

This study addresses the challenges of world model maintenance and risk perception under visual failure by proposing a modality-agnostic abductive framework that treats missing acoustic evidence as latent causal cues to trigger inference. Integrating microphone array localization, Doppler approach rate extraction, and Neyman-Pearson detection, the method effectively sustains the world model and outputs calibrated risk warnings during visual occlusion. Experimental results demonstrate that the system provides alerts 1.7 seconds earlier in blind curve scenarios while reducing false alarm rates by 42%. Furthermore, it achieves a risk perception accuracy exceeding 0.87 under visual degradation, significantly enhancing safety in non-line-of-sight environments.

Abstract

A structured world-state (entities, relations, context, and predictive cues) is designed to preserve prediction-critical content when perception degrades, but it presumes observations to populate it; when the primary visual modality is occluded or degraded, those observations may be missing. We address how to sustain the world model from a complementary modality by treating the absence of expected co-evidence as evidence of a hidden cause. The abductive framework is modality-agnostic; this article instantiates it acoustically. A microphone-array front-end estimates the bearing of engine and tire sources and extracts approach-rate evidence (Doppler when a stable tone exists, a broadband looming readout otherwise); the event "signature present, visual co-evidence absent" then triggers abductive inference of a hidden road user, emitting a calibrated risk advisory rather than a control command. Recoverability of the hidden state is analyzed as an identifiability question separating shared from modality-unique information, and cueing is cast as Neyman-Pearson detection under an explicit false-alarm budget. On real occluded-approach recordings at blind junctions, the method warns a mean 1.7 seconds before line-of-sight entry, matches the sustained-window variant of the published acoustic baseline's detection rate with 42% fewer false alarms, localizes to 3.4 degrees median once in view, is well calibrated (expected calibration error 0.034), and keeps hazard awareness above 0.87 under staged vision degradation that collapses a vision-only channel to 0.03. We also measure the method's limits: calibration transfers to an unseen junction almost losslessly, the signature classifier does not, and moving-ego noise is the binding deployment constraint.

Problem

• World Models
• Cross-Modal Perception
• Abductive Reasoning
• Risk Perception
• Visual Degradation

Innovation

• Cross-Modal Abductive Reasoning
• Evidence of Absence
• World Model Sustainability
• Acoustic Risk Perception
• Neyman-Pearson Detection