CUSP: CUSUM-Governed Survival Hazard Alarms at the Perception Onset for Off-Road Navigation

2026-10-06

AI Summary

This study addresses the limitation of existing off-road navigation methods that rely on intervention logs, which struggle to detect danger critical points as early as human perception in real time. We propose a model-agnostic runtime danger alert mechanism that, for the first time, models early human judgment rather than delayed interventions. By annotating "perception onset points," our approach combines a visual danger head with discrete-time survival analysis to learn the transition moment from safe to hazardous states, and introduces a CUSUM-based cumulative risk alarm strategy. Evaluated across five unseen sites, the proposed mechanism successfully detects 85 dangerous events, substantially outperforming the best baseline, which identifies only 26. These results demonstrate its effectiveness in recognizing potential hazards within conventional signal blind spots.

Abstract

Off-road navigation exposes a robot to potentially hazardous terrain en route. Although learning-based navigation uses safety supervision to choose which path to drive, it provides no runtime alarm when the robot following that path is heading into danger. Such an alarm must be learned from field logs, where human intervention preempts the failure and the failure itself is therefore never observed. The human judges driving unsafe early but typically intervenes only once failure is clearly near, so the intervention marks that judgment late. That earlier judgment is what a runtime alarm must detect, yet no prior intervention-supervised method has targeted it. To address this problem, we introduce CUSP (CUSUM-governed Survival model of the Perception onset), a model-agnostic runtime hazard alarm that learns this moment from intervention-terminated logs. "Cusp" is a word for the point at which one state is about to turn into another, and the moment we target is exactly such a cusp: the point at which safe driving turns unsafe in a human's judgment. We call this point the perception onset and annotate it separately from the intervention. A visual hazard head is trained on the annotated onset with a discrete-time survival objective so that driving with and without an onset both supervise the head, and a CUSUM accumulates the predicted onset risk into alarms. We evaluate CUSP at five unseen sites, two autonomous and three teleoperated, with 142 events and every method tuned to the same rate of ten false alarms per hour. CUSP detected 85 events compared to 26 for the best of nine adapted baselines, and the margin comes from hazards to which every signal in the navigation model is blind.

Problem

• off-road navigation
• runtime hazard alarm
• human intervention
• perception onset
• survival analysis

Innovation

• Survival Analysis
• Perception Onset
• CUSUM
• Off-Road Navigation
• Hazard Alarm