Title: DNAS: a driver nighttime assistance system using rear-view smartphone
Authors: Chunmei Ma; Haigang Gong; Nianbo Liu; Chao Song; Ming Liu
Addresses: School of Computer Science and Engineering, University of Electronic Science and Technology of China, Sichuan Province, 610054, China ' School of Computer Science and Engineering, University of Electronic Science and Technology of China, Sichuan Province, 610054, China ' School of Computer Science and Engineering, University of Electronic Science and Technology of China, Sichuan Province, 610054, China ' School of Computer Science and Engineering, University of Electronic Science and Technology of China, Sichuan Province, 610054, China ' School of Computer Science and Engineering, University of Electronic Science and Technology of China, Sichuan Province, 610054, China
Abstract: Owing to the poor visibility and improper behaviour, driving at nighttime is much more dangerous. In this paper, we present a cost-effective driver nighttime assistance system (DNAS), which runs on Android smartphones. In DNAS, smartphones are placed on the rear windshield, for periodically capturing road condition by the camera sensor, and warning the drivers of dangerous speeding or tailgating event. To achieve this function, firstly, DNAS uses the bright vehicle headlights and their dimension to determine the following vehicles from the captured images. Then, it analyses their temporal and spatial characteristics in the successive frames. Finally it works out the safety critical events to send out warning sounds. We evaluate the effectiveness of DNAS based on two different real driving routes. The results show that DNAS, on average, is able to detect speeding and tailgating with an accuracy of 87.7% and 83.3% recall, respectively.
Keywords: DNAS; driver nighttime assistance system; nighttime driving; smartphone.
DOI: 10.1504/IJAHUC.2017.086262
International Journal of Ad Hoc and Ubiquitous Computing, 2017 Vol.26 No.2, pp.104 - 114
Received: 20 Oct 2014
Accepted: 07 May 2015
Published online: 04 Sep 2017 *