Bowen Geng

Bowen Geng is a PhD candidate at Rutgers University, specializing in Computer Vision, LiDAR Sensing, Connected Autonomous Vehicle Systems, and Traffic Flow Theory.

Focus

  • Vision-based car-following models and driver behavior models
  • Computer vision for intelligent transportation and mobility applications
  • Roadside sensing → Connected Vehilce messaging (MAP/SPaT/BSM/PSM)

About

I’m a PhD candidate in Civil Engineering at Rutgers University. My work bridges computer vision, roadside sensing (LiDAR/camera), and connected-vehicle communications to enable reliable real-time traffic perception and analytics.

Research

Visual Driving Rule (VDR)

A car-following model derived from 2D image plane, with theoretical stability analysis and simulation.

Traffic Signal Phase & Timing from Roadside Camera

Traffic light detection + adaptive color features over time, robust to illumination, glare, and partial occlusion.

Birds Eye View (BEV) Reconstruction from dashcam

Transforming front-view dashcam footage into top-down road layouts for vehicle localization and scene understanding.

Driving 3D Digital Twin

Reconstructing dynamic road scenes in 3D to support digital twin visualization, traffic analysis, and mobility research.

Projects

Real-time Messaging (MQTT / ETX)

Authenticated publish/subscribe data products for MAP/SPaT/BSM-like streams over cellular networks.

TLS • Protobuf • Paho MQTT

Android Dashcam App

Developed an application to record dashcam video and corresponding mobile sensor data.

Java • Android • Mapping SDK

Publications

  1. Microscopic Driver Behavioral Fundamental Diagram Analysis Using Computer Vision Techniques: Implications for Human Driving Responses to Visual Stimuli TRB Annual Meeting, 2025 Presentation (TRB 2025) PDF Code
  2. New Brunswick Innovation Hub Smart Mobility Testing Ground Jin, Peter J. et al., FHWA-NJ-2024-003, 2024 ROSA/National Transportation Library
  3. The Development of the Digital Twin Platform for Smart Mobility Systems With High-Resolution 3D Data Jin, Peter J. et al., CAIT-UTC-REG45, 2021 ROSA/National Transportation Library

Contact

Email: bg605@scarletmail.rutgers.edu
LinkedIn: /in/bowengeng