Multisensor System for Vehicle Detection of 13 Classes (081-FH4), Phase 2b
Project Information
Traffic sensors for detection of motorcycles approaching signals have not been sufficiently robust. False negatives indicating the absence of motorcycles when motorcycles are present can induce motorcyclists to run the red indication, creating a potentially hazardous situation. This project will extend the capabilities of vision-based motorcycle sensors by adding a stereo vision component; note that the policy office is also interested in this project because it is currently difficult to characterize vehicle miles of travel for motorcycles due to the lack of sensors, which accurately count and classify motorcycles as separate vehicles. This project hopes to accomplish the following: (1) Verify whether stereo vision and sound emission measurement performs well for motorcycle detection. (2) Identify critical gaps regarding the collection and processing of stereo vision data, and what hardware techniques or algorithms might be used to add to the stereo motorcycle detection techniques to add robustness to the classification and separate motorcycles from bicycles and mopeds. (3) Conduct laboratory and field tests of improved hardware and algorithms.
Goals:
The key project objective is to detect motorcycles with advanced sensors.
- Operations
- FY 2002-2022 / Operations / Automation and Connectivity
An advanced sensor can reliably detect motorcycles.
AMRP = Annual Modal Research Plan