Overview
Research explored the optimal timing for proactive in-cabin agents to communicate with passengers in fully automated vehicles (AVs). The objective was to determine how communication should adapt to an event's priority and the passenger's current activity. The study compared two distinct communication policies: an event-triggered (ET) policy and a context-sensitive (CS) policy, evaluating their impact on communication appropriateness, perceived interruption, and passenger trust.
Research Context
Proactive in-cabin agents are designed to assist passengers in comprehending the behavior of automated vehicles. However, a potential challenge lies in the risk of over-communication, where relaying every ride event might lead to unnecessary interruptions for the passenger. This study addresses the need for a more nuanced communication strategy to balance information delivery with passenger comfort and engagement.
Approach
A mixed-methods, within-subject study design was employed, involving 41 participants. Each participant rode as a passenger in a virtual reality (VR) simulated fully-automated vehicle. The study directly compared two communication policies:
- Event-Triggered (ET) Policy: Under this policy, the in-cabin agent communicated immediately upon the occurrence of every event.
- Context-Sensitive (CS) Policy: This policy allowed the agent to select one of three communication modes for each event: Immediate, Delayed, or Silent. This selection was informed by the event's priority and the passenger's activity.
The study collected data on communication appropriateness, perceived interruption, and perceived trust to assess the efficacy of each policy.
Findings
- Communication Appropriateness: The context-sensitive (CS) policy increased communication appropriateness compared to the event-triggered (ET) policy.
- Perceived Interruption: The CS policy substantially reduced perceived interruption among passengers.
- Perceived Trust: No significant difference in perceived trust was observed between the ET and CS policies. However, baseline dispositional trust was found to differentiate communication preferences among participants.
- Key Considerations for Selective Communication: The research highlighted several factors as crucial for effective selective in-cabin communication. These include the consequence of an event, the passenger's activity state, the continuing information value of a communication, and the passenger's need for confirmation regarding vehicle actions or events.
Why This Matters
The findings suggest that a context-sensitive approach to in-cabin agent communication can enhance the passenger experience in automated vehicles by making information delivery more appropriate and less intrusive. Understanding the specific factors that influence the need for and timing of communication—such as event consequence and passenger activity—is crucial for designing more effective and user-friendly AV interfaces.