pacific nw transportation consortium region 10 Pactrans

Fusion of Airborne and Terrestrial Sensed Data for Real-time Monitoring of Traffic Networks

PI: Sameh Sorour (UI), samehsorour@uidaho.edu, ORCID: 0000-0002-3936-7833

Co PIs: Mohamed Hefeda

AMOUNT & MATCH: $40,000 from PacTrans; $40,000 Match

PERFORMANCE PERIOD: 8/16/2018 – 8/15/2020

STATUS: Completed

CATEGORIES: Real-time Sensing Data, Traffic Network

RESEARCH PROJECT HOT SHEET:

UTC PROJECT DOCUMENTATION:

FINAL PROJECT REPORT:

PROJECT DATA:

DESCRIPTION: This research project will build upon the project team’s current work on real-time object recognition and event extraction from LiDAR scans using UAS on-board processing, while also incorporating a suite of complementary sensors and recent advances in transfer learning and geometry matching techniques. Advanced algorithms will be employed to fuse the objects recognized from multiple sources, increasing the accuracy and robustness of traffic network monitoring and detection of features/events of interest. A key benefit of the terrestrial sensors will be to enable the construction of recognition maps, even when UAS data acquisition is infeasible due to regulatory and/or logistical considerations. The proposed solution relies on the collection of recognized objects from the same site by different sensing sources, which are then transmitted to a fusion center. This fusion center will apply transfer learning and geometry matching techniques to both create correspondences between these detected objects (i.e., determine which object corresponds to which other object across the different data sets) and add undetected objects/zones by some of the sources into their proper positions. The final integrated recognition map will provide much richer and more robust information to traffic network controllers, thus enabling data-driven optimization and efficiency for transportation networks.