Autonomous vehicles: Reliability of their perception of the world around them and the role of human driver

Abstract

According to the vision of technology promotors, in a fully autonomous vehicle design a driver need not be in the loop even in a complex driving environment. Recognizing the technological, legal, social and economic constraints, this vision may not become reality for many years into the future. However, a “cognitive vehicle” provides a realistic step towards full automation. A cognitive vehicle is an intelligent vehicle designed to have a seamless nondistracting interface with the driver and for reasons of liability and cost, the driver remains “in the loop”. As a part of this vehicle’s design, an adaptive driver assistance system, among many other features, can have the longitudinal and transverse control capability. The driver has the option to turn on the active safety feature and can override the automation feature to make difficult decisions. On the other hand, active feature takes over driving if the driver is highly distracted or disabled.