Deep-Learning Based Trajectory Forecast for Safety of Intersections with Multimodal Traffic
With the convergence of computation, communication, sensing, and visualization into ever smaller and cheaper devices, the United States is entering in a new era of more efficient, safer and more affordable transportation. The recent emergence of novel augmented reality technologies in particular offer a formidable opportunity to improve the safety of traffic for all road users. In this proposal, the objective is to investigate the potential of augmented reality, in conjunction with deep-learning based image processing, for traffic safety at intersections, considering all possible modes of transportation. A major problem of a safety system is to offer reliable, timely warnings, with a low false detection rate. This somehow requires the prediction of the future trajectories of the users, which is the primary focus of this proposal. In this project, the goal is to investigate how deep learning can be used to detect motion cues, and estimate over a short time horizon the future path of road users (depending on their transportation modes) using real-time video data. Once estimated, these paths will be used as part of an augmented-reality based system, where information about potential conflicts is displayed in the field of view of all road users, through smart glasses.
Language
- English
Project
- Status: Completed
- Funding: $90821
-
Contract Numbers:
69A3551747133
-
Sponsor Organizations:
Center for Advanced Multimodal Mobility Solutions and Education
University of North Carolina, Charlotte
Charlotte, NC United States 28223Office of the Assistant Secretary for Research and Technology
University Transportation Centers Program
Department of Transportation
Washington, DC United States 20590 -
Managing Organizations:
University of North Carolina, Charlotte
Department of Civil and Environmental Engineering
9201 University City Boulevard
Charlotte, NC United States 28223-0001 -
Project Managers:
Fan, Wei
-
Performing Organizations:
University of Texas at Austin
Austin, TX United States 78712 -
Principal Investigators:
Claudel, Christian
- Start Date: 20171001
- Expected Completion Date: 20190930
- Actual Completion Date: 20190930
Subject/Index Terms
- TRT Terms: Communication; Computer algorithms; Forecasting; Image processing; Intersections; Machine learning; Multimodal transportation; Traffic safety; Vehicle trajectories; Virtual reality
- Subject Areas: Data and Information Technology; Highways; Operations and Traffic Management; Safety and Human Factors;
Filing Info
- Accession Number: 01652820
- Record Type: Research project
- Source Agency: Center for Advanced Multimodal Mobility Solutions and Education
- Contract Numbers: 69A3551747133
- Files: UTC, RIP
- Created Date: Dec 3 2017 8:28PM