TL;DR
This study investigates how familiar drivers' eye fixation patterns differ from unfamiliar drivers at urban intersections. It reveals that familiarity leads to reduced attention to the front and increased attention to the sides and rear, potentially explaining higher accident risks for familiar drivers, though not specifically rear-end collisions.
What Problem Does This Paper Address?
While route familiarity is known to influence driving behavior and accident risk, its specific impact on drivers' visual attention, particularly at complex urban intersections, has been underexplored. Existing research often relies on simulations or focuses on general driving, leaving a gap in understanding how familiarity alters eye fixations in real-world urban intersection scenarios. This study aims to bridge this gap by examining how familiar and unfamiliar drivers allocate their visual attention at these critical points.
Methods
The researchers conducted an on-road experiment in Changsha, China, using an eye-tracking system (Tobii Pro Glasses 2) to record drivers' eye fixations. Participants were categorized as "familiar" (driving the route at least once a month) or "unfamiliar" (driving the route at most once a year). The experiment involved driving a 9 km route with 9 intersections. Data was collected 150 meters before and 50 meters after the lane stop line at these intersections. Visual fixation measures, including total duration, average duration, and number of fixations, were analyzed across ten Areas of Interest (AOIs) representing different parts of the driver's view. Statistical analysis, specifically independent-samples T-tests, was used to compare the fixation behaviors between the two groups.
Datasets
The study utilized data collected from 24 participants (12 familiar, 12 unfamiliar) driving on a specific 9 km route in Changsha, China. This route was selected based on accident data from 2014-2019, highlighting it as an accident black spot. No pre-existing datasets were used for training or evaluation; the study generated its own dataset through the on-road experiment.
Key Insights and Ideas
- Familiarity shifts attention: Contrary to a simple "less attention" hypothesis, familiarity doesn't necessarily mean less overall visual processing. Instead, it leads to a reallocation of attention, with familiar drivers looking less at the immediate front and more towards the sides and rear.
- Mental workload implications: Unfamiliar drivers exhibit higher processing times and mental workloads for the front of the vehicle, suggesting they are actively processing more information in this critical area. Familiar drivers, conversely, show increased sampling rates for the rear and sides, indicating a different attentional strategy.
- Distraction paradox: The study did not find evidence supporting the idea that route familiarity directly leads to increased distraction at urban intersections. This challenges a common assumption about familiar drivers.
- Explaining accident risk: The observed decrement in familiar drivers' eye fixation strategies, particularly their reduced focus on the front, could partially explain their higher accident risk. However, this doesn't fully account for the higher incidence of rear-end accidents among familiar drivers, suggesting other factors are at play.
Key Results
- Frontal Fixations: Both groups focused heavily on the front (RF, FL, FR), but unfamiliar drivers showed significantly longer fixation durations on the right front (RF) (p = 0.042), indicating a higher mental workload.
- Side and Rear Fixations: Familiar drivers had a significantly higher sampling rate for the right mirror (RM) (p = 0.009) and spent more time looking at the driver's sides (LW, RW) and rear (LM, RM, ReM) compared to unfamiliar drivers. Specifically, familiar drivers checked the left window (LW) 28.5% more frequently (p = 0.039).
- Processing Time: Familiar drivers showed a reduced sampling rate and processing time for their front view but an increased rate for their sides, particularly the left window (p = 0.007).
- Mental Workload: Unfamiliar drivers experienced higher mental workloads for the right front (RF), while familiar drivers had higher workloads for the left window (LW) and "other" areas (OT).
Limitations and Future Work
The authors acknowledge several limitations:
- Limited Scope: The study focused solely on the effect of route familiarity on visual fixations at urban intersections in Changsha. It did not examine the interaction between driving experience and route familiarity, nor did it consider different types of intersections or turning maneuvers.
- Generalizability: The findings are specific to the urban environment of Changsha, China, and may not generalize to other driving contexts or cultures.
Future work suggested includes:
- Investigating Accident Mechanisms: Further research should explore the reasons behind familiar drivers' higher involvement in rear-end accidents, potentially by examining drivers' visual scanning strategies, physiological signals, and overall driving behavior.
- Broader Factors: Future studies should incorporate driving experience and intersection type to provide a more comprehensive understanding of visual fixation differences.
- Turning Maneuvers: The impact of different turning maneuvers on visual attention and route familiarity should be investigated.
- Brain Activity and Fatigue: Linking eye movements to brain activity and identifying mental fatigue could offer deeper insights into accident mechanisms.
- Warning Systems: The effectiveness of warning systems and different warning methods on driving safety and economy should be explored.
- Intelligent Vehicles: Incorporating route familiarity into autonomous vehicle (AV) research can help anthropomorphize AVs and improve their interaction with human drivers.
Summary updated