COVERT: Cognitive Internet of Vulnerable Road Users in Traffic

Kamat, Vineet R.; Menassa, Carol C.; Zhou, Xiaoshan · 2025 · ROSA P / United States. Department of Transportation. University Transportation Centers (UTC) Program

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study addresses the critical safety challenge of predicting pedestrian road-crossing intentions, particularly when pedestrians are occluded or stationary, which limits the effectiveness of existing vision-based and reactive sensing methods. Current approaches, including cooperative vehicle sensing and deep learning trajectory models, struggle with sudden appearances from blind spots or the inability to detect intent before movement begins. To bridge this gap, the authors propose a proactive framework using wearable electroencephalogram (EEG) signals to detect motor planning prior to physical action. The research aims to enable autonomous vehicles to anticipate pedestrian movements through intelligent vehicle-to-everything (V2X) systems, thereby reducing collision risks for vulnerable road users. The researchers conducted experiments where participants, embodied as visual avatars, interacted with simulated traffic scenarios involving varying volumes, marked crosswalks, and traffic signals. EEG data were collected using a wearable headset and processed through a multi-stage computational framework. First, Power Spectral Density and time-frequency analyses (using Morlet wavelet and multitaper methods) decomposed signals into theta, alpha, beta, and gamma bands. Functional connectivity was assessed using the Weighted Phase Lag Index to mitigate volume conduction artifacts. To model the temporal evolution of cognitive processes, the team constructed a Gaussian Hidden Markov Model (HMM) to decompose EEG sequences into four latent cognitive microstates: perception, risk assessment, timing determination, and movement initiation. Finally, motor readiness was predicted using a K-nearest Neighbors algorithm combined with Dynamic Time Warping on sliding window datasets. The results demonstrated that high-beta oscillations in the frontocentral cortex served as a strong predictor of motor readiness. This specific neural marker achieved an Area Under the Curve (AUC) of 0.91, providing an anticipatory lead window of approximately one second before the participant initiated physical crossing movement. The HMM successfully identified the transitions between cognitive microstates, validating the hypothesis that distinct neural patterns underlie the perception-action loop in road-crossing decisions. These findings indicate that EEG signals can reliably capture the internal decision-making process before it manifests as observable motion. The significance of this work lies in its shift from reactive to proactive pedestrian-vehicle interaction. By detecting motor intentions via brain activity, the proposed framework allows autonomous systems to respond earlier than current sensor-based methods, potentially preventing accidents caused by occlusion or sudden crossings. The study establishes a novel method for modeling fast-evolving neural modulation in real-world decision-making contexts. Furthermore, the framework is adaptable to other human-robot interactions, such as bicyclist safety or industrial robotics, offering a pathway for seamless collaboration in dynamic, connected traffic environments.

Key finding

High-beta oscillations in the frontocentral cortex predicted pedestrian motor readiness for road crossing with an Area Under the Curve of 0.91 and approximately one second of anticipatory lead time.

Methodology

lab_experiment

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• anticipation
• pedestrian behavior perception
• vru facing ehmi

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data
• Methodological Resource: tool software
• Theoretical Contribution: computational model