534
Views
0
CrossRef citations to date
0
Altmetric
Peer-Reviewed Journal for the 27th International Technical Conference on the Enhanced Safety of Vehicles (ESV)

Silent failure detection in partial automation as a function of visual attentiveness

ORCID Icon, ORCID Icon, &
Pages S88-S93 | Received 12 Aug 2022, Accepted 20 Nov 2022, Published online: 02 Jun 2023

Figures & data

Figure 1. NADS high-fidelity motion-based simulator (left), AutoDrive display (right).

Figure 1. NADS high-fidelity motion-based simulator (left), AutoDrive display (right).

Figure 2. Gaze AOIs: front, rearview mirror, left window and mirror, phone task.

Figure 2. Gaze AOIs: front, rearview mirror, left window and mirror, phone task.

Figure 3. Scan path sequences for one event.

Figure 3. Scan path sequences for one event.

Figure 4. Four clusters in 30 second sequences describe driver behaviors. (a) disengaged, (b) mostly disengaged, (c) split evenly, (d) engaged.

Figure 4. Four clusters in 30 second sequences describe driver behaviors. (a) disengaged, (b) mostly disengaged, (c) split evenly, (d) engaged.

Figure 5. Hidden Markov Models for each cluster. Circles denote hidden states. State compositions show emission probabilities for each gaze AOI. (a) disengaged, (b) mostly disengaged, (c) split evenly, (d) engaged.

Figure 5. Hidden Markov Models for each cluster. Circles denote hidden states. State compositions show emission probabilities for each gaze AOI. (a) disengaged, (b) mostly disengaged, (c) split evenly, (d) engaged.

Figure 6. Simulated glance response times during the silent failure event. Large dots show actual glance response times. Driver behavior prior to the event is color coded.

Figure 6. Simulated glance response times during the silent failure event. Large dots show actual glance response times. Driver behavior prior to the event is color coded.
Supplemental material

Supplemental Material

Download Zip (5.4 MB)