Currently, there are commercially available vehicles that include features capable of providing Level 2, Partial Driving Automation as defined by SAE International. Research on the use and performance of the systems that these vehicles employ in natural settings is needed to help clarify the systems’ potential benefits. The Naturalistic Study of Level 2 Driving Automation Functions (L2 NDS) project described herein has generated practical data to support the understanding of the use of automated lateral and longitudinal control functionality by evaluating a subset of currently available advanced technologies as drivers experience them during daily use.

The objective of the L2 NDS project was to investigate, through a naturalistic driving study, real-world driver interaction with commercially available driving automation systems. Ten vehicles equipped with both lateral and longitudinal automated features were instrumented and loaned to participants for a 4-week period. A total of 120 drivers were recruited over a 14-month data collection period. Each study vehicle was equipped with Virginia Tech Transportation Institute’s NextGen Data Acquisition System, which continuously records video of the both the driver and the roadway, as well as vehicle data and automated lateral and longitudinal control activations. These data were used to analyze driving automation system use and driver performance during the study.

Focus area 1 investigated System Performance, including overall use of the features. Participants drove 216,585 miles, with 70,384 miles driven with both lateral and longitudinal control features active. Focus area 2 investigated Driver-System Interaction and involved a review of driver behaviors during driving automation system use, specifically the prevalence of non-driving tasks. Drivers were observed engaging in non-driving tasks, but these were not related to feature use. Focus area 3 investigated Driver Performance, which was measured by drivers’ responses to Request to Intervene (RTI) alerts generated by the driving automation systems. Driver behavior was consistent with active driving/supervision of the automated features;​ drivers were receptive to RTI alerts. No RTIs were associated with any safety-critical events (i.e., crashes and near-crashes). In total, 5 minor crashes (no injury or visible damage) and 66 near-crashes were observed across the entire data set. No statistical relationship was observed between safety-critical event rates and feature activation level. Focus area 4 investigated Driver Engagement, which includes subjective feedback obtained from participants. Participants reported that they were generally comfortable and felt safe using the features, with self-reported trust increasing over the course of the study.