Human Driver Performance Assessment based on HiLCPS Concept

Abstract

The paper presents an approach for human driver performance assessment from the view of his/her abilities or limitations. The idea is based on the Human-in-the-loop Cyber-Physical Systems (HiLCPS) concept which is briefly described in the initial part of this paper. A cybernetic approach describing a human driver as a controller in cooperation with a controlled system represented by a simulated driving task is there used. In this context, a car-driving simulator based on Unreal Engine 4 was developed. A human driver controls a simulated drive via an input HCI (Human-Computer Interface) device, i.e., a joystick or a steering wheel with pedals, exploiting the perception of the current situation in the visualized scene. That simple HiLCPS system provides data representing human behavior during a specific task under defined conditions. Based on the acquired data we are able to identify the behavioral model as well as other quantities characterizing the attitude to control. These parameters then serve for describing a current state of a human driver and his/her performance which is demonstrated in the last section.