Ethnography: Informing Personas and Usability Test Scenarios to Evaluate external-Human-Machine Interface (eHMI)
Sponsor: Nissan Research Center - Silicon Valley (NRC-SV)
Course: ANTH 232 - Advanced Ethnographic Applications
Team: 3 Applied Anthropology (M.A.) Graduate Students
Contribution: Research Design \ Data Collection and Analysis \ Preparation and Presentation of Deliverables
Ethnographic research was conducted in collaboration between NRC-SV and the Applied Anthropology M.A. program at SJSU to inform the development of pedestrian personas and usability test scenarios to evaluate external-Human-Machine Interface (eHMI).
Some vehicle-pedestrian interaction scenarios require driver-to-pedestrian communication to safely negotiate the interaction
Fully-Autonomous Vehicles (AVs) eliminate the role of the driver and need for a human operator, disrupting a critical vehicle-pedestrian communication channel and introducing new pedestrian safety risks
Investigate vehicle-pedestrian interactions and communication
Identify Vulnerable Road Users (VRUs) and critical vehicle-pedestrian interaction scenarios that require driver-to-pedestrian communication to safely negotiate the interaction
What are the behaviors and interactions of suburban bicyclists (both single and group cyclists)?
How do suburban bicyclists (both single and group cyclists) and drivers negotiate their interactions?
Literature Review \ Observation \ Interview \ Survey
Observation Quota Sample: (N = 10-Hours)
10-hours of raw video footage were obtained to approximate data saturation across two user categories: (1) Single Cyclist (n = 5) and (2) Cyclist Group (n = 5). In user research, data saturation refers to the stage in data analysis where no new information is discovered by the researchers, signaling the ability to cease additional data collection and analysis. A sample size of five (n = 5) in each user category is typically required to approximate data saturation.
Interview Quota Sample: (N = 20)
A total of 20 representative users were solicited, screened, scheduled for a 60-minute, one-on-one interview. Participants were selected to approximate data saturation across the following user categories:
Pedestrian behavior is significantly influenced by individual, social, and environmental factors
Pedestrian age, group size, and dynamic, environmental factors have large effect sizes
Vehicle-to-Pedestrian (V2P) communication systems (i.e., vehicle horn and turning signals) safely negotiate most micro-and-macro interactions
Driver-to-Pedestrian communication (e.g., non-verbal or verbal communication) is critical to safely negotiate many micro-and-macro interactions
Observational footage captured a macro vehicle-pedestrian interaction between tens-of-vehicles collectively signaling and communicating their intent to turn right at a busy intersection and hundreds of disregarding bicyclists simultaneously cycling straight through, creating a dangerous traffic jam of frustrated drivers.
Drivers largely ignored and blocked by disregarding bicyclists become visibly frustrated, yelling at cyclists and even playing 'chicken' with pedestrians
Driver-to-pedestrian communication captures the attention of a cyclist who immediately after begins to swerve, signaling the rest of the group to a complete stop
Hundreds of bicyclists stop behind the pedestrian crossing line (and behind the signaling cyclist), allowing several drivers to make a right turn and clearing the intersection of vehicular traffic
Organizers and Volunteers
"We have rules that participants are supposed to follow for their own safety. Unfortunately, we can't really enforce them"
- Organizer, M, 40
Cyclists and Participants
"It's easier to get away with stuff if you're riding with a big group. You just have to watch out for angry drivers"
- Bicyclist & Participant, M, 19
"Sometimes you just have to drive around them. Avoid them"
- Driver, F, 30
Survey (N = 112)
82% of respondents believe groups of cyclists have the highest priority over the road use
70% believe vehicles have a higher priority over the road use than single cyclists
23% reported using hand signals ALL or MOST of the time when riding alone
7% reported using hand signals ALL or MOST of the time when riding in groups
By triangulating and converging on the analyzed observational, interview, and survey data, the following insights were concluded:
Driver-to-pedestrian communication is critical to safely negotiate many micro-and-macro vehicle-pedestrian interactions
Developers must account for the social engineering aspects of fully autonomous (i.e., artificial intelligence) systems, including autonomous vehicle to pedestrian communication.
The following deliverables were presented to the NRC-SV Design Team and ANTH 232 Advanced Ethnographic Applications course at SJSU:
The following personas represent VRUs in vehicle-pedestrian interactions.
Daredevil Dave, M-15
Dave's bike is his main source of transportation
Loves riding with his friends and causing mayhem
Does not appreciate authority
Commuter Carole, F-30
Carole rides her bike to-and-from work every day
Takes all necessary precautions and can be overly cautious
Believes that following the rules will keep her safe, despite the flow of traffic
The Fun Family
Each member of the "Fun" family owns a bike
Always wear helmets and practice safe behaviors
Family Rule: Never ride alone!
The following experience flow maps a critical vehicle-pedestrian interaction scenario that requires driver-to-pedestrian communication to safely negotiate the interaction.
A short-film storyboard was produced to help designers develop empathy for VRUs, accounting for them in the design of AV2P communication systems.
This research provided insights into V2P communication systems and driver-to-pedestrian communication in micro-and-macro vehicle-bicyclist interaction scenarios that require driver-to-pedestrian communication to safely negotiate the interaction. Additionally, critical vehicle-bicyclist interaction scenarios were developed to test future AV2P communication systems. Further research on VRUs and critical interaction scenarios is needed to develop viable AV2P systems, including testing with representative users (VRUs) and across critical interaction scenarios.
M.S., Human Factors/Ergonomics Thesis Project:
This project inspired and informed the development of Autonomous Vehicle-Pedestrian Interaction and Communication Simulator (AV-PICS), a traffic simulator for Human Factors (HFE), pedestrian research. The computer application was developed for an SJSU HFE M.S. project to evaluate the usability of Autonomous Vehicle-to-Pedestrian (AV2P) communication systems.