United States Department of Transportation
ROSA P serves as an archival repository of USDOT-published products including scientific findings, journal articles, guidelines, recommendations, or other information authored or co-authored by USDOT or funded partners.
As a repository, ROSA P retains documents in their original published format to ensure public access to scientific information.
i
Evaluating Reliability of Tesla Model 3 Driver Assist Functions
-
2020-10-01
[PDF-2.15 MB]
Details:
-
Alternative Title:Evaluating the Reliability of Tesla Model 3 Driver Assist Functions
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:Collaborative Sciences Center for Road Safety ; United States. Department of Transportation. Federal Highway Administration ; United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:SEP 2019- SEP 2020
-
Corporate Publisher:
-
Abstract:The number of vehicles with advanced driver-assist systems (ADAS) that can simultaneously perform automated steering and acceleration is increasing, with the expectation that these capabilities will be available in low-cost models as well as luxury cars. These systems require human drivers to be alert and available at all times in case they need to take over. Thus, the interface between the human and the machine is of critical importance in ADAS-equipped vehicles. Formal testing and guidance for development of such ADAS systems is limited. The goal of this effort was to assess between- and within-ADAS-equipped vehicle variation in four scenarios involving the interface between a human driver and an ADAS system. These scenarios were: (1) Assessing driver-monitoring system performance during automated highway driving; (2) Alerting a distracted driver of unexpected road patterns during automated driving; (3) Assisting a distracted driver in response to an inadvertent lane departure; and (4) Initiating driver handover to a distracted driver when the vehicle can no longer confidently operate. Given that Teslas can face a range of potentially demanding environments, a Tesla Model 3S was the test platform. The results suggest that the performance of the computer vision systems was extremely variable, and this variation was likely responsible for some, but not all, of the delays in alerting a driver whose hands were not on the steering wheel. Performance was not consistent for a single car, in that a car would perform the best in the most challenging driving scenarios (navigating extreme curves while the driver ignored takeover requests), but performed the worst on seemingly simpler scenarios like detecting a lane departure. These results suggest that practitioners need to develop a richer set of tests that capture within- and between-car variability, and they also indicate that software upgrades through over-the-air updates may induce latent issues that lead to safety issues.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
