pacific nw transportation consortium region 10 Pactrans

An Airborne Lidar Scanning and Deep Learning System for Real-Time Event Extraction Control Policies in Urban Transportation Networks

PI: Christopher Parrish (OSU), christopher.parrish@oregonstate.edu, ORCID: 0000-0002-2681-0090

Co PIs: Sameh Sorour (UI), Ahmed Abdel-Rahim (UI), David Hurwitz (OSU)

AMOUNT & MATCH: $180,000 from PacTrans; $180,000 Match

PERFORMANCE PERIOD: 8/16/2017 – 8/15/2019

STATUS: Completed

CATEGORIES: UAVs, Urban Transportation Networks, Deep Learning

RESEARCH PROJECT HOT SHEET:

UTC PROJECT DOCUMENTATION:

FINAL PROJECT REPORT:

PROJECT DATA:

DESCRIPTION: The project team is currently investigating the capability to provide transportation and mobility solutions driven by real-time data generated from UAS using lidar and event identification through deep learning. Specific project tasks include: 1) developing optimal UAS-based lidar acquisition methodologies (payloads, sensor settings, and processing strategies) for transportation network scanning; 2) designing, implementing, and testing a deep learning algorithm that can extract features from the UAS lidar data, and 3) developing guidelines for state DOTs and other transportation agencies on the technical and operational requirements for UAS-based lidar data integration. The OSU project team recently integrated a Velodyne Puck lidar system and OxTS xNAV direct-georeferencing system on a DJI S1000 remote aircraft and have conducted test flights under an FAA-issued Certificate of Authorization (COA). Next steps will include working with ODOT to identify project sites to scan with the UAS-based lidar and transmitting the data to the UI project partners for implementing and testing the deep learning algorithms.