NeuroFaceLab: A new framework for passengers analysis in autonomous driving

摘要

Driverless cars are seen as one of the key disruptors in the nexttechnology revolution . However, user acceptance needs to bedealt with in order for autonomous vehicles to be successfullyintroduced to the market. Contrary to traditional vehicleswhere the acceptability of a model is basically determined by theacceptability of the vehicle by its driver, for autonomous cars it isnecessary to analyze the acceptability for a group of passengers.Although “user acceptance” is an abstract term that has beenredefined multiple times based on need and purpose, our goal inthis research is to quantitatively measure emotional states of thepassengers from the classification of their facial expressions andtheir correlation with attention and meditation indexes computedfrom their brain waves. Focusing on above goal, this workintroduces a new framework to study passenger’s acceptance of anautonomous vehicle. The framework (called "NeuroFaceLab")allows analyzing the emotional states of the passengers of anautonomous vehicle. The system has 2 modules: a data acquisitionmodule and an analysis module. In the data acquisition module, avideo of approximately 20 minutes of a driving in a simulator isshown and a video of the users is obtained synchronously duringthe user’s observation of the video in combination with their brainwaves. In the analysis module the position of the users' face isdetected and tracked to obtain their facial expressions. In thispreliminary report, the main problems in the application of thesystem to detect the emotional states with 7 users are reported andthe correlation of their facial expressions with their brain waves isdiscussed.In the