Enhancing Driver Awareness Using See-Through Technology 2018-01-0611
This paper presents a real-time application of see-through technology using computer vision (e.g., object detection) and Vehicle-to-X (V2X) communication (e.g., Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I)). Each access point (AP) was connected to Chattanooga’s fiber optics internet, supporting a data transfer rate up to 10-Gbps. Using a 5Ghz frequency, vehicular communications were set up with a seamless handover for transferring real-time data. Two web cameras acting as clients were mounted on the windshield of two of three vehicles to send image data to the offsite server. Using multi-threaded programming, both image feeds were processed simultaneously. Once the server received the images, it performed an object recognition algorithm on each image using a convolutional neural network (CNN). Post- identification, the images from the second vehicle were sent and overlaid dynamically to the third vehicle’s image. This repetitive overlapping of images allowed the third vehicle to “see-through” the second vehicle in real-time. This experiment was showcased during the US Ignite Smart Cities Summit in June 2017 to emphasize the benefits of drivers being able to “see-through” the car in front to make more intelligent decisions when passing a vehicle, stopping for a pedestrian, or seeing an upcoming detour due to construction before the view is within their line of sight. Using V2X communication with computer vision gives the driver a higher level of awareness and allows better decision making in the case of a roadway conflict, ultimately increasing the level of safety on our roadways.
Citation: Thompson, R., HU, Z., Cho, J., Stovall, J. et al., "Enhancing Driver Awareness Using See-Through Technology," SAE Technical Paper 2018-01-0611, 2018, https://doi.org/10.4271/2018-01-0611. Download Citation
Author(s):
Rebekah L. Thompson, Zhen HU, Jin Cho, Jose Stovall, Mina Sartipi
Affiliated:
University of Tennessee at Chattanooga
Pages: 9
Event:
WCX World Congress Experience
ISSN:
0148-7191
e-ISSN:
2688-3627
Related Topics:
Fiber optics
Neural networks
Real-time data
Imaging and visualization
Vehicle drivers
Vehicle to grid (V2G)
Vehicle to vehicle (V2V)
Internet
Windows and windshields
Mathematical models
SAE MOBILUS
Subscribers can view annotate, and download all of SAE's content.
Learn More »