Communication Network Architectures for Driver Assistance Systems
Open Access
- 16 October 2021
- Vol. 21 (20), 6867
- https://doi.org/10.3390/s21206867
Abstract
Autonomous Driver Assistance Systems (ADAS) are of increasing importance to warn vehicle drivers of potential dangerous situations. In this paper, we propose one system to warn drivers of the presence of pedestrians crossing the road. The considered ADAS adopts a CNN-based pedestrian detector (PD) using the images captured from a local camera and to generate alarms. Warning messages are then forwarded to vehicle drivers approaching the crossroad by means of a communication infrastructure using public radio networks and/or local area wireless technologies. Three possible communication architectures for ADAS are presented and analyzed in this paper. One format for the alert message is also presented. Performance of the PDs are analyzed in terms of accuracy, precision, and recall. Results show that the accuracy of the PD varies from to depending on the resolution of the videos. The effectiveness of each of the considered communication solutions for ADAS is evaluated in terms of the time required to forward the alert message to drivers. The overall latency including the PD processing and the alert communication time is then used to define the vehicle braking curve, which is required to avoid collision with the pedestrian at the crossroad.
Keywords
This publication has 15 references indexed in Scilit:
- Machine learning for internet of things data analysis: a surveyDigital Communications and Networks, 2018
- Pedestrian Movement Direction Recognition Using Convolutional Neural NetworksIEEE Transactions on Intelligent Transportation Systems, 2017
- YOLO9000: Better, Faster, StrongerPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2017
- Automatic Incident Detection in Intelligent Transportation Systems Using Aggregation of Traffic Parameters Collected Through V2I CommunicationsIEEE Intelligent Transportation Systems Magazine, 2017
- SSD: Single Shot MultiBox DetectorPublished by Springer Science and Business Media LLC ,2016
- Towards Real-Time Object Detection on Embedded SystemsIEEE Transactions on Emerging Topics in Computing, 2016
- Early Detection of Sudden Pedestrian Crossing for Safe Driving During Summer NightsIEEE Transactions on Circuits and Systems for Video Technology, 2016
- Discriminative Unsupervised Feature Learning with Exemplar Convolutional Neural NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2015
- A Hybrid Approach for Automatic Incident DetectionIEEE Transactions on Intelligent Transportation Systems, 2013
- Detection of Sudden Pedestrian Crossings for Driving Assistance SystemsIEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2011