Research 01
Proactive Crash Risk Prediction and Prevention
Can risk be identified before a crash? I combine connected-vehicle trajectories, traffic states, and driving behavior to support timely, context-sensitive prevention.
Explore researchResearch
From traffic data to safety evidence.
My research connects four questions: how to anticipate roadway risk, explain its behavioral and physical mechanisms, evaluate vehicle technologies, and understand safety-critical events through multimodal evidence.
Research 02
Crash Mechanisms, Human Factors, and Roadway Design
How do people and roads shape safety? I examine injury mechanisms, driver perception, vulnerable road users, and context-sensitive roadway design.
Explore research
Research 03
Autonomous Driving, ADAS, and CAV Safety
How can driving technologies respond safely to hazards? I study safety recovery, cyber-physical resilience, and the roadway conditions affecting ADAS performance.
Explore research
Research 04
Multimodal Traffic Scene Understanding
What happened, why, and when should a warning occur? I connect video, trajectories, and structured signals to interpret and evaluate safety-critical traffic events.
Explore researchMethodological foundation
Different questions. Connected methods.
My work brings together causal inference, Bayesian and random-parameter modeling, spatiotemporal analysis, and multimodal machine learning. Methods are chosen around the safety question and the available evidence.
Related publicationsResources & benchmarks
Data, protocols, and safety evaluation
Preprint
SAVeD: A First-Person Social Media Video Dataset for ADAS-equipped vehicle Near-Miss and Crash Event Analyses
A multimodal dataset project for first-person ADAS near-miss and crash event understanding.
Research detailsSubmitted
SAVeD: A Deployment-Oriented Protocol for Candidate-Event ADAS Warning-Evaluating Video, Structured Signals, and Timing Sensitivity in Crash-Risk Escalation
Examines a deployment-oriented protocol for ADAS candidate-event warning, combining video, structured signals, and timing-sensitive evaluation.
Research detailsSubmitted
ALARM: A Local Action Safety Boundary Benchmark and Model for End-to-End Pre-Crash Safety Recovery
Investigates a local-action safety-boundary benchmark and model for pre-crash safety recovery.
Research detailsSubmitted
MSC-MTSim: A Novel Simulation Platform for Multi-Class Cyberattacks on Mixed Traffic with High-Fidelity Dynamics and Lane-Change
Investigates a mixed-traffic simulation platform for multiple cyberattack classes, vehicle dynamics, and lane changes.
Research detailsContact
Let's connect.
I welcome collaborations on proactive traffic safety, human factors and roadway design, automated vehicle safety, and multimodal traffic scene understanding.
chenzhu.wang@ucf.edu
Smart & Safe Transportation Lab
University of Central Florida
Orlando, Florida, USA