Blog

Disclaimer: The information provided in this blog post is intended for educational purposes and not legal advice. For specific legal inquiries, please consult an attorney. This information is intended for the sole purpose of reconstructing accidents and injuries and not to serve as a risk management resource. 

LISKE's lead Human Factors engineer, Dr. Miao Song, specializes in matters related to automobiles, drivers, and other road users. In one of his peer-reviewed published papers “Is Driving Automation Used as Intended? Real-World Use of Partially Automated Driving Systems and their Safety Consequences”, Dr. Song investigated driver interactions with driving automation. This was prompted by the introduction of advanced driver assistance systems featuring SAE Level 2 automation into the vehicle marketplace

This investigation was prompted by concerns that drivers might not fully understand the capabilities and limitations of driving automation systems (DASs), which promise to fundamentally alter the driving experience by providing automated control over lateral and longitudinal vehicle movements.

Recent tests raise concerns about the real-world reliability of lower-level Driving Automation Systems (DAS) and drivers' misuse of such systems, exemplified by incidents involving Level 2 DASs like Tesla's AutoPilot. This study analyzed data from 50 participants driving vehicles equipped with Level 2 DASs over 12 months, covering 684,931 miles, to understand their usage patterns and performance. Researchers examined objective data and participants' opinions about these systems, collected from a post-study questionnaire targeting the usefulness and usability of the DASs. The research aimed to answer questions about the frequency of safety-critical events involving DASs; the system's responsiveness during these events; instances of unintended DAS use; user perceptions of DAS usefulness, and usability; and instances where DASs failed to meet driver expectations.

It was found that among 235 safety-critical events (SCEs), DASs were involved in 20% of them. Longitudinal control features like adaptive cruise control (ACC) were used more frequently than lateral control features such as lane keep assist (LKA). Interestingly, in 27% of vehicles, engagement of one feature required another to be activated first. Out of the 47 SCEs with DAS use, the system issued an alert or controlled the vehicle in 29.8% of cases. However, DASs were not intended to address the threat in 14.9% of events, resulting in no alert issued. Additionally, in 13% of SCEs, the systems neither reacted nor warned the driver. A thorough examination revealed that 47 safety-critical events (SCEs) that involve DASs were misused in 57% of safety-critical events. This misuse frequently involved drivers engaging in secondary tasks or using the systems in situations not suitable for highway driving.

Furthermore, participants generally viewed DASs positively for their usefulness and ease of use, with a preference for longitudinal over lateral control features. However, tests revealed a greater preference for ACC over LKA among participants. Additionally, some reported manually turning off DASs. Participants felt more comfortable engaging in secondary tasks when DASs were active. Familiarity with DAS increased with usage, though fewer participants expressed trust in the system. Some experienced issues with DAS functionality, such as insufficient braking or acceleration. These findings suggest mixed sentiments towards DASs' performance and reliability, with specific concerns regarding braking and disengagement.

Want to speak with an Expert? Reach out to our team at experts@liskeforensics.com or call us at 1-888-674-3508.

Sources:
o  Kim, H., Song, M., & Doerzaph, Z. (2022). Is Driving Automation Used as Intended? Real-World Use of Partially Automated Driving Systems and their Safety Consequences. Transportation Research Record, 2676(1), 30-37. https://doi.org/10.1177/03611981211027150