@article{BjelicaITSM2023a,
title = {Reliability of Self-Driving Cars: When Can We Remove the Safety Driver?},
author = {Milan Z. Bjelica and Bojan Mrazovac},
url = {https://ieeexplore.ieee.org/document/10056924},
doi = {10.1109/mits.2023.3244271},
issn = {1939-1390},
year = {2023},
date = {2023-07-01},
journal = {IEEE Intelligent Transportation Systems Magazine},
volume = {15},
number = {4},
pages = {46-54},
publisher = {Institute of Electrical and Electronics Engineers (IEEE)},
abstract = {Self-driving cars and other vehicles are being increasingly demonstrated, released, and deployed on roads. Yet only a handful of special-case vehicles have obtained a permit to go fully driverless, and our passenger cars still require drivers at the steering wheel to correct the car if something goes wrong. Safety drivers, which operate autonomous vehicles in testing phases, upon occasions were unable to correct the faulty behavior of their vehicles, inducing accidents and, unfortunately, casualties. This article proposes a method to analyze the reliability of autonomous vehicles applying the classical reliability theory and software reliability growth models (SRGMs). This method is then applied to real-world data to get some predictions on what is really needed to safely remove the driver and go fully driverless.},
howpublished = {M21},
keywords = {automotive, functional safety, reliability},
pubstate = {published},
tppubtype = {article}
}
Self-driving cars and other vehicles are being increasingly demonstrated, released, and deployed on roads. Yet only a handful of special-case vehicles have obtained a permit to go fully driverless, and our passenger cars still require drivers at the steering wheel to correct the car if something goes wrong. Safety drivers, which operate autonomous vehicles in testing phases, upon occasions were unable to correct the faulty behavior of their vehicles, inducing accidents and, unfortunately, casualties. This article proposes a method to analyze the reliability of autonomous vehicles applying the classical reliability theory and software reliability growth models (SRGMs). This method is then applied to real-world data to get some predictions on what is really needed to safely remove the driver and go fully driverless.