Fundamental Generation Systems

2Reliability of Autonomous Systems

2.1. Introduction

Autonomous systems have the ability to act on their own, in order to carry out actions that are necessary to achieve predefined objectives in response to stimuli which, in robotics, for example, come from sensors. There is the notion of intelligent control and the reliability of the system. Intelligent control involves algorithms, linguistics and mathematics applied to systems and processes (Cardon and Itmi 2018). Let us take the example of the autonomous vehicle, where some autonomous car models are impressive in their reliability and ability to detect obstacles and avoid them. Detection, however, remains too random for the moment: unfavorable conditions (poor visibility, obstacle not recorded in the database) can cause such systems to fail. Current examples of these vehicles include the Tesla X and Tesla S, BMW 5 series, Volvo S90, Audi A8 and Mercedes S class, with BMW offering different levels of autonomy, from the first level offering no assistance to the last level of autonomy, where the vehicle drives by itself. According to several experts, the growth of the Industry 4.0 concept and autonomous systems are the challenges of tomorrow. The expected performances of these systems are the perception of their environment and the ability to interpret different situations by algorithms, based on artificial intelligence or machine learning. Among the systems most used in the industry, there are radars, ultrasonic and inertial ...