Software Reliability of Complex Systems Focus for Intelligent Vehicles

• Published in: Vehicle and Automotive Engineering pp. 309 - 321
• Main Authors: Schuster, György, Tokody, Daniel, Mezei, Imre János
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2017; Springer International Publishing
• Series: Lecture Notes in Mechanical Engineering
• Subjects: Automotive technology; Complex systems; Intelligent vehicles; Metals technology / metallurgy; Safety critical system; Software reliability
• ISBN: 9783319511887; 3319511882
• ISSN: 2195-4356; 2195-4364
• DOI: 10.1007/978-3-319-51189-4_28

Using software became a part of our everyday life, in the last few decades. Software is widely used in areas, such as national defence, aeronautics and astronautics, medicine or even transport. There are 100 million lines of codes in a modern high-end car’s engine control unit. In comparison, the Space Shuttle needs 400 000, the F22 fighter jet needs less than 2 million, the Boeing 787 airplane needs 14 million and the Facebook needs more than 60 million lines of codes to function. Even a smaller error can lead to devastating consequences in safety-critical systems, such as those operating in vehicles. There have been several examples in recent years, when an automotive recall was necessary due to dangerous software, and there were cases when these errors presumably caused fatal accidents. Definition of software reliability is the error-free working probability of software for a specified period of time under well-defined environment. Usage of software is inevitable. It can be found in every vehicle to control almost everything. Therefore software can be considered as a critical success factor and it has a strong effect on the reliability of the whole system. The software systems are getting more and more complex. Known fact is a more complex system has more possibility to have errors. The most difficult problem is that the traditional methods of reliability cannot be used. For example fatigue and wearing of mechanical parts or features of lubricant systems can be calculated quite well, since we have enough prior knowledge on their features. Unfortunately, in case of software systems this knowledge is missing. This paper deals with the question of software reliability. In the first part it lists the problems and the second part gives some mathematical issues to calculate working probability.