Driver Assistance Using Automatic Understanding of the Driving Environment

Source Organization:
Principal Investigator:Prof. Martial Hebert, Dr. Luis Navarro-Serment
PI Contact
Project Manager:
Funding Source(s) and Amounts Provided (by each agency or organization):
Total Dollars:$75,000
Agency ID/Contract/Grant Number:
Start and End Dates:
Project Status:Complete
Subject Categories:Safety, transportation, pedestrians and bicyclists, research
Abstract:An intelligent driving system must be cognizant of two basic elements to recommend effective and accurate actions: 1) the environment surrounding the vehicle, and 2) the expected movements of other objects. In this project we investigate advanced concepts to develop techniques for building internal models of the vehicle’s surrounding environment (including both static and dynamic objects), which can be used to assist drivers by warning them of risky or dangerous situations and recommending preventive actions. Our overall approach is to combine: 1) machine learning and scene understanding techniques with onboard sensors to classify and model the environment, 2) information from external data sources--such as maps--to boost the modeling accuracy and to provide additional contextual information, and 3) reinforcement learning techniques to model relationships between observed behavior and features in the environment, to predict the trajectories of moving objects.
Describe Implementation of Research Outcomes (or why not implemented):
Impacts/Benefits of Implementation (actual, not anticipated):
Project URL: