Description
Blind spots are a well known problem in traffic causing both bodily and material damages. BlindSpots aims to eliminate the problem by utilizing augmented reality (AR) to make the affected areas of the vehicle “see through”. An ecosystem of cameras are used to gather information outside the vehicle and streams it to the AR glasses. When the driver looks at the direction affected by a blind spot, the system allows the driver to see through the surface of the vehicle, thereby effectively mitigating the blind spot. Functionally, this can be seen as similar to the way one uses rear windows in an ordinary car. During the project several stakeholders have emerged and have been engaged in discussions about the solution including: Danish Road Safety Agency (Færdselsstyrelsen), CODAN, 3F and DTL. These stakeholders have all shown a common interest and excitement for the upcoming product. The tests performed have been conducted with a low fidelity prototype built from VR glasses, an android smartphone and a webcam. The main experiments have been carried out with a Ford Ka where blind spots have been simulated by blinding out windows. Additionally, the system was tested on the DTU garbage truck and the driver in collaboration with smart campus. The results from the 30+ participants has shown a benefit in the perception of drivers, specifically indicating a reduction in the error rate when using the prototype. The project having shown positive benefits will carry on into a master’s thesis using the Microsoft HoloLens. During the next 6-7 months the low fidelity prototype will be replaced with a high fidelity prototype aiming to showcase and test a fully functional system.
The contact provided by Jens Dahlstrøm to the DTU garbage truck drivers allowed us to test the experience in a close-real life setting. Testing the experience with end-users.
Image example of how the augmented reality looks in practice.
A comparison of the total errors in testing (for 30 users), using mirrors and AR respectively.
Data - Are they available for others to use?
Performance benchmarking, error rates, user experience data. The data is personal and is
thus not publicly available.
Contact
Dan Roland
s134837@student.dtu.dk
Tanja Lind Hansen
s173341@student.dtu.dk
Thomas Bjelbo Thomsen
s170452@student.dtu.dk
or contact Smart Campus to hear more.