This book is not intended to be an encyclopedia on autonomous and/or connected vehicles. Its sole purpose is to explain the types, choices, operation, properties, and architectures of the networks that can be used in autonomous vehicles, depending on a great many external parameters. Thus, one part of the book (the earlier chapters, essentially) details these parameters, with the aim of quantifying their technical implications in concrete terms – for example, in terms of the choice of network topology, datarate, latency, level of security, performance, compliance with norms, standards and regulations, etc.
In this book, various worlds collide – mainly, the world of automobiles, its connected domains, and numerous entities in the worlds of electronics, mechanics, and communication. All of these disciplines have their own specific sets of vocabulary, their own ways of being and acting, design methods, marketing techniques, and commercial approaches, which are generally very different – and this is perfectly normal.
Often, the thinking in electronics can be viewed as linear and Cartesian, taking one step at a time in logical succession. However, in the automotive world, it becomes much more serpentine and interlinked with other factors, because everything reacts to (and influences) everything else, and often we need to view the product that the end client wants as a contiguous whole, rather than a collection of subsets.
Before setting out on this long journey, let us make two specific points.
On the subject of autonomous and connected vehicles, the Internet offers hundreds of articles (some better than others), presenting some complex and marvellous theories, all sorts of varied and vast future markets, fabulous forms of encryption, etc. As we are not fond of unproductive redundancy, we have focused solely on subjects about which there are not as many articles – i.e. the down-to-earth, day-to-day of this domain, offering a concrete and technical discussion of the vast range of applications and designs. The aim, in so doing, is to guide readers, to overlook nothing, and to avoid the pitfalls that may be encountered in the process of designing and implementing secure, autonomous, and connected vehicles. It is all very well to talk about such things, giving speeches and lectures and demonstrations (as we have seen and heard many times). However, to concretely and physically realize a connected solution for commercial purposes, and to sell it in large quantities at a sensible and reasonable price, is far better. Otherwise, it would be as well to do nothing, and forego the unnecessary fuss. This book describes the procedures that must be observed to avoid the usual pitfalls in a project, and facilitate the transition from the virtual to the real and concrete world. Thus, we propose an approach based on due consideration of the technical, financial, ergonomic, etc. standards, rather than on false promises.
On this subject, in late 2018, Bernard Favre, an expert at Inria and the former Head of Research at Volvo-Renault Trucks and of the LUTB Transport and Systems industrial research program, wrote the following. “Autonomous vehicles are a highly complex technology, in which it is probably harder to bring artificial intelligence to bear than in any other application. In no other sector is the technology faced with such a diverse range of situations. At present, we are in the full throes of innovation “in the lab.” As yet, there is no real proven market …. The number of tests that automakers require in order to validate an autonomous vehicle’s performances is soaring. They include physical experiments in real conditions, and digital simulations. … Having a certain amount of experience of the disparity between what automakers’ projections and announcements say about when new technologies will be available, and when they actually become available for commercialization (for various reasons: maturity, regulations, market acceptability, cost, real performance, etc.), I fear that autonomous vehicles will be no different to what I have seen time and time again in my career”. He concludes by projection that “autonomous cars will be operating on private circuits by 2025. In relation to autonomous vehicles on public/open roads, it is likely to be 2040. …”
This is a view which we, the authors, have long shared.
This, then, is the explicit aim of this book, which should therefore remain on your coffee tables as a reference until 2035 at least – and that should be enough!