A Bayesian Game in the Human-Robot Society: The interaction between pedestrian and two types of autonomous vehicle
The speaker will be Daehyun Yoo, who is visiting us from University of Siena, where he is doing his PhD in Economics. He will give a talk on “A Bayesian Game in the Human-Robot Society: The interaction between pedestrian and two types of autonomous vehicle”
Main content
Speaker
Daehyun Yoo
Title
A Bayesian Game in the Human-Robot Society: The interaction between pedestrian and two types of autonomous vehicle
Abstract
Autonomous vehicles (AVs) have the potential benefits to improve traffic safety and efficiency. However, there is an ethical dilemma in the context of protecting different stakeholders, such as pedestrians and passengers of AV. People prefer other people’s AV to protect pedestrians at the expense of sacrificing the passengers. But they prefer their own AV to prioritize protecting passengers. This inconsistent moral preference of participants in AV challenges Google, Apple and other car manufacturers about what types of self-driving cars to develop and produce. Few research consider the moral gap between expectation and reality to program AVs. In this article, I propose a new approach to categorize insiders and outsiders of AV as two different types of human participants in the AV decision process. Two types of AV are represented by it. Then, I use a Bayesian game model to analyze the strategic interactions with pedestrians. It shows that the existence of two different types of AV leads to two pure Nash Equilibria. Because pedestrians are uncertain about AV’s type, the model resolves pedestrian’s moral hazard behavior. This model also suggests that the adoption of different two types of AV improves transportation efficiency in mixed traffic. At the same time, it may give car manufacturers an incentive to advertise their operating ethical principles and enhance transparency.