Toward Informed AV Decision-Making: Computational Model of Well-being and Trust in Mobility (2505.14983v1)

Abstract: For future human-autonomous vehicle (AV) interactions to be effective and smooth, human-aware systems that analyze and align human needs with automation decisions are essential. Achieving this requires systems that account for human cognitive states. We present a novel computational model in the form of a Dynamic Bayesian Network (DBN) that infers the cognitive states of both AV users and other road users, integrating this information into the AV's decision-making process. Specifically, our model captures the well-being of both an AV user and an interacting road user as cognitive states alongside trust. Our DBN models infer beliefs over the AV user's evolving well-being, trust, and intention states, as well as the possible well-being of other road users, based on observed interaction experiences. Using data collected from an interaction study, we refine the model parameters and empirically assess its performance. Finally, we extend our model into a causal inference model (CIM) framework for AV decision-making, enabling the AV to enhance user well-being and trust while balancing these factors with its own operational costs and the well-being of interacting road users. Our evaluation demonstrates the model's effectiveness in accurately predicting user's states and guiding informed, human-centered AV decisions.

Computational Model of Well-being and Trust in Autonomous Vehicle Decision-Making

The paper introduces a sophisticated computational framework designed to enhance autonomous vehicle (AV) decision-making by integrating human cognitive states, specifically well-being and trust, into the process. This approach stems from the increasing social presence of AVs and the necessity to align their operations with human needs, particularly within shared mobility contexts.

In this paper, the authors present a Dynamic Bayesian Network (DBN) model capable of inferring the cognitive states of an AV user and other road users. These inferred states are subsequently utilized to influence the AV's decisions in a manner that promotes user well-being and trust. The DBN effectively integrates variables such as the AV user's intention, trust levels, and well-being, along with the interactions and well-being of other road users.

Empirical Study and Model Evaluation

The authors conducted an empirical paper involving interactions between self-driving scooters and delivery robots, focusing on strategic accommodating actions such as yielding versus non-yielding. The paper's data were used to refine the parameters of the DBN model, which enables the inference of latent states based on observable factors during AV interactions. The evaluation showcased the model's capacity to predict user states with notable accuracy: well-being predictions reached 77%, trust predictions 67%, while intention predictions were highly precise at 95%.

The effectiveness of this model in capturing dynamic interplay among cognitive states crucially informs AV decisions, providing a foundation for systems that advocate for smoother and more human-centered autonomous mobility.

Implications for Autonomous Systems

This research provides valuable insights into the trajectory of AV systems, emphasizing user-centric design and interaction. The dynamic nature of well-being and trust, as modeled in this paper, underscores the importance of incorporating nuanced human factors into autonomous decision-making processes. Ultimately, this framework proposes a more informed approach to AV decision-making that prioritizes human well-being, aligning technological advancements with societal needs.

Future Directions

As autonomous systems continue to evolve, future research could expand upon this foundation by exploring longitudinal data for richer temporal dynamics, incorporating diverse mobility modes, and investigating broader scenarios with multiple road users. This model's adaptability could also encourage the exploration of other domain-specific variables and contexts, propelling advancements in AI that respect human-centric principles.

In conclusion, this paper contributes to the field of autonomous vehicle research by presenting a robust framework capable of understanding and integrating key human cognitive states into decision-making processes, paving the way for AV systems that are not only technologically advanced but also socially considerate.