Quickest Detection of Abnormal Vehicle Movements on Highways
Document Type
Conference Proceeding
Publication Date
7-1-2019
Publication Title
FUSION 2019 - 22nd International Conference on Information Fusion
Keywords
Autonomous vehicles, collision avoidance, collision detection, likelihood ratio test, Page test, road safety, self-driving cars
Abstract
The quest to develop self-driving vehicles remain an active research topic. A self-driving vehicle on a highway employs numerous sensors to track the state of its surrounding vehicles. Considering the proximity of surrounding vehicles, it is critical to detect their unusual maneuvers as quickly as possible, especially when autonomous vehicles operate among human-operated traffic. In this paper, we present an approach to quickly detect lane-changing maneuvers of a nearby vehicle. The proposed algorithm is based on the optimal likelihood ratio test, known as Page test. The proposed approach is presented in the form of a novel process model to be employed by autonomous vehicles. In addition to traditional states, such as position and velocity, the proposed process model adds two additional states of a surrounding vehicle being monitored: the lane-index (LIDX) and the lane-change-index (LcIDX). Then we present an approach to keep these two indices up to date in the quickest possible manner through the proposed Page test based algorithm.
ISBN
9780996452786
Recommended Citation
Wang, Jingyu; Dhanapal, Ravindra Kumar; Ramakrishnan, Priyadharshini; and Balasingam, Balakumar. (2019). Quickest Detection of Abnormal Vehicle Movements on Highways. FUSION 2019 - 22nd International Conference on Information Fusion.
https://scholar.uwindsor.ca/computersciencepub/117