Energy Consumption and Autonomous Driving

Among the trends that are going to shape the automotive industry in the coming years, autonomous driving stands out as having the potential to completely change the automotive industry as we know it. While analysts may still debate the pace of change, the current state of autonomous driving technology or the power dynamics between incumbents and new entrants, there is no longer a debate over if autonomous driving is going to happen, but when. For traditional players in the automotive industry, this means they have a series of strategic questions to answer that will determine the path to the autonomous driving future and their roles in it. This paper is an excerpt of a Roland Berger Strategy Consultants study recently published on the topic [1].

Modern vehicles consist of many interconnected, software-based IT components which are tested very carefully for correct functional behavior to avoid safety problems, e.g. the brakes suddenly stop working. However, in contrast to safety testing systematic testing against potential security gaps is not yet a common procedure within the automotive domain. This however could eventually enable a malicious entity to be able to attack a safety-critical IT component or even the whole vehicle. Several real-world demonstrations have already shown that this risk is not only academic theory [1]. Facing this challenge, the paper at hand will first introduce some potential automotive security attacks and some important automotive security threats. It then explains in more detail how to identify and evaluate potential security threats for automotive IT components based on theoretical security analyses and practical security testing.

V2X communication is making its way to series-production to fulfil the vision of zero-fatalities safety and enhanced mobility for lower emissions. V2X benefit is heavily depending on the technology penetration rate. The penetration will increase on two conditions: technology is cheap and can be introduced quickly. Following experience gained from several design cycles, Autotalks obtains a deep understanding of optimizing V2X system cost, fitting any vehicle architecture or retrofit device. The presented optimization methods cover pre-integrated software, optimal module partitioning and optimized chipset solution. Examples of addition to several vehicle architectures will be demonstrated. The paper will show that cost-effective V2X solution, added at low-risk and effort, is available now.

Car connectivity will contribute to coping with future challenges such as growing urban populations, traffic jams, pollution and road fatalities. This paper exposes in brief NXP’s view, technology and recent experiences in past and ongoing projects that will accelerate the deployment of V2X.

Advanced driver assistance systems (ADAS) are becoming ubiquitous, and the safety of these systems is more important than ever. Engineers are incorporating new technologies—including different sensing modalities, data processing and fusion algorithms, as well as enhanced control and automation systems—leading to increased system complexity. In order to manage the complexity of these systems and address the issue of safety via extensive testing, it is vital to have an integrated environment so that different technological components can be incorporated and tested at a system level. We present such a solution in this paper. The proposed framework is based on Model-Based Design, enabling vision algorithms and control strategies to be developed using the same platform. Furthermore, software-in-the-loop (SIL) testing can be achieved via automatic C/C++ code generation. To close the loop at a system level, the environment must also be modeled. Environmental conditions include the weather conditions, the type of sensor used which affects the input data for the vision algorithms, as well as the vehicle dynamics model. This aspect is also addressed in this paper. We focus on lane-keeping as an example to illustrate different aspects of the framework.

This paper looks at the implications on the interfaces between the different domain levels and vehicle control system for highly automated driving (HAD). Scenarios or scenes provide the basis for this new development process. Using these scenes, a set of function specifications can be compiled that cover all possible use cases. Implementation of this type will in future demand to an even greater extent that a function is first developed without considering implementation before it is then distributed over the target system with different controllers. And, moreover, higher demands are placed on the physical accuracy of the data provided by the domain management system compared to today’s assistance systems.

Autonomous driving requirements regarding safety, redundancy and performance lead to an increasing number of on-board heterogeneous equipments (sensors, actuators and processing units). Systems with the highest levels of autonomy must handle a large number of real-life situations, some of them being quite challenging. Interactions between sensors, perception, planning, control and actuators are fundamental to ensure operational efficiency while providing both proactive safety and timeliness execution. To address these constraints, appropriate design patterns are required to implement complex and robust decisional architectures. As a matter of fact, current system vehicle standards do not address these design neither patterns nor their underlying complexity. Based on various experiences in defense and civilian domains, we will present a critical review of ongoing standard developments and propose a generic design pattern to qualify autonomous vehicles architectures.

In the “Internet world”, the car is rapidly becoming one of the most connected elements in our everyday’s life. Our research topic deals more specifically with drivers who connect their smartphones to carpool with strangers. In fact, smartphones communicate crucial data for eco-mobility, such as the number of “empty seats travelling” [1], available for potential passengers. Thanks to the GPS, 3/4G networks and “dynamic carpooling” applications, the car stands out as the new “public-private” transport. This innovative service of dynamic carpooling develops in a lightning way. In this article, we decided to present a state of the art, which details successively the history of the practice, the technical components of the service, the issues and needs for a sufficient critical mass of users, the benefits and limits of the system. Then, we describe various examples of incentives, updated during the deployment of large-scale studies to encourage the practice to a larger number of users.

Legal considerations have to be taken into account in the development of new technologies that concern safety, security and more general the society. This article describes the legal framework from different angles including product liability and criminal with a particular attention on the Vienna convention, French law and international evolutions.

Autonomous driving becomes real. For the first time ever the insertion of artificial intelligence into self-propelled machines, which can now operate themselves is likely to make human being as a driver redundant. Machines have the legal status of manufactured objects or instruments, but are the legal rules applicable to objects able to capture the consequences of the machine’s ability to move freely particularly? The distribution of responsibility and product liability, technical standards and obviously the insurance matter are to be shaped in this respect. Besides, a special focus shall be given on data and cyber security which need to protect data autonomy and privacy to reassure potential consumers.

Testing of highly automated vehicles has new challenges with respect to the questions to answer, the test cases, and the testing procedures. Main questions arise from the fact that highly automated vehicles are required to achieve high levels of availability and effectiveness of the vehicle functions; after all, their performance has to be compared to the performance of human drivers. Testing of such vehicles requires international consensus on the required level of safety and on the metrics to be applied. The main challenges for such testing are discussed and some new approaches are presented.

To efficiently develop and validate such systems a generic and integrated solution providing Sensor Simulation, Virtual and Rapid Prototyping and Real Time Bypassing is needed. This paper proposes such a solution, with a comfortable editor with various predefined objects e.g. for ACC, LDW, Break-Assist, autonomous parking/driving to support easy configuration for the different validation views; access to automotive buses like CAN, CAN-FD, FlexRay, Ethernet; high speed internal access to the ECU data e.g. through Microcontrollers Data Trace Interfaces; Video Streams acquisition through various camera interfaces including Ethernet/BroadR-Reach; GPS Position data. The solution also includes a real-time bypassing and rapid prototyping execution platform. The solution works independent from the ECU supplier implementation and can be also used to compare various ECU vendors.

Amongst automotive functionalities which generate brakings like AEBS, it is necessary to assess the consequences of an inopportune braking in the presence of a following vehicle. In order to meet this requirement, a specific methodology was developed. This methodology uses data recorded on open roads and the ISO 26262 fundamentals. The main risk for ADAS function in particular for automatic braking, is to trigger a braking which can be perceived as inopportune, whereas the vehicle is followed by another vehicle which does not react quickly enough and collides. In order to define the failure rate criterion, we use the relation between the “pre-existing risk” and the product Exposure (E) by Controllability (C) for each Severity (S). By using data from accidentology, it is make it possible to evaluate the pre-existing risk and using a recorded database, compute the product [E × C]. Finally, the methodology allows to define the failure rate requirement necessary to drive without safety-related incident in order to validate the function.

Since the 80s, the building of learn and test data bases for learning-based systems (i.e. neural networks) had to cope with problems of picking representative examples and measuring the generalization/the score of the system. And of course, real open world applications cannot be fully tested. It seems that artificial vision-based ADAS now discover the same question, and then, may use the same solutions, involving the same methodology (A.G.E.N.D.A.), using design of experiments and data analysis tools.

Automotive industry is facing challenges to reduce CO₂ emissions. A promising approach consists in anticipating the road profile and the upcoming dynamic events like traffic lights. V2X technologies enable this anticipation and allow CO₂ emission reduction as well as traffic flow improvement. This topic has been addressed in the framework of the French public funded project Co-Drive, with the use of traffic lights data broadcasted to vehicles. One of the developed functions by Valeo within the Co-Drive project is a Green Light Optimal Speed Advisory (GLOSA) system. This system coaches the driver to adapt his vehicle speed in order to safely pass the next traffic lights during the green phase. It allows reducing stop times and unnecessary accelerations in urban traffic situations and therefore saving fuel and reducing CO₂ emissions. Indeed, state-of-the-art studies showed the great potential of GLOSA systems in terms of CO₂ emission reduction and traffic flow improvement with different approaches. Here we present a description of the GLOSA system that has been implemented on a Valeo demonstration car. It has been tested with promising results and got very positive feedbacks from customers and public authorities. Next developments of V2X communication like green phase for emergency vehicle approaching, traffic light violation signal, or adaptive routing will allow further improvements in CO₂ emission reduction, safety and comfort.

The AVL software package “upgrade-E” enables the predictive calculation of the expected speed, elevation and tractive power requirements of an unknown driving route. The development platform mainly accesses freely available data formats such as Open Street Map (OSM) and SRTM altitude profiles. The predicted driving route provides a plethora of optimisation possibilities for classical vehicle and powertrain functions. The prototype version of the software is implemented in AVL’s electric vehicle Coup-e 800 and runs on a conventional 7-inch tablet PC with appropriate data gateways to the vehicle’s CAN bus.

PSA Peugeot Citroën has unveiled in 2013 a new hybrid powertrain concept—Hybrid Air-described as a disruptive hybrid system that makes a major step to lower emissions and fuel economy. With this system during city driving cycle, vehicle operates in Air Mode from 60 to 80 % of the time. As thermal engine operation is resultant of this on and off mode, the alternator is mainly shut off; in consequence, the lead-acid battery has strong cycling increase and can be quickly damaged. Besides, it’s not suitable to power the electrical equipment during such a long time with only battery voltage (lower than 14 V). To fit those requirements, PSA Peugeot Citroën has developed an electrical system called “SPRESSO” performing two major functions: quick storage and energy management. It can handle with high efficiency, the charge and the discharge of a high-capacity electric double layer capacitor (EDLC), minimizing spare stored energy and makes it possible to supply electricity to all vehicle equipment during Air Mode. SPRESSO is compatible with standard 12 V architecture (alternator and lead-acid Battery) and offers a long life and cost effective electrical system to Hybrid Air equipped vehicles.

Advanced Driver Assistant Systems (ADAS) are constantly improving in performance and fitment rates in modern cars. While in the past parking aids have been a driver’s selection for a price premium, today most vehicles offer park distance control as a standard configuration. When developing driver assistant systems it is important to select the most suitable technology. Parking aids for instance typically rely on ultrasonic or camera sensors. In premium cars a combination of both sensor technologies is implemented to improve the system reliability and at the same time increase the user’s benefit. A single CMOS imager camera cannot measure the distance of an object but an image is a good method to interact with the driver. Hence, the fusion of sensors results to more sophisticated systems. The availability of multiple assistant systems in a vehicle will ultimately lead to autonomous driving. Today, autonomous or semi-autonomous driving is used for low-speed parking maneuvers. Mainly undefined legislation framework is slowing down a wider adoption rate. Recently US government has lowered the requirements for self-driving vehicles. Can autonomous cars be sold as soon as 2015? This paper reviews different sensing technologies for driver assistant systems, highlights the benefits and necessity of sensor fusion and presents an optimized approach for utilizing infrared time-of-flight (TOF) sensors to implement cost efficient, reliable, scalable and precise environmental sensing for driver assistance and comfort applications.

Smart Power Technologies, first introduced at the end of the 80s in the industry, cover today all the range of Automotive applications from Power Train to Safety, Body Electronics and Infotainment, with important macro trends of evolution and innovation led by Energy Saving. The paper describes the evolution of “BCD Smart Power” technologies for Automotive applications, looking for combined technology platforms were digital content, high voltage and current are present in the single chip. Advanced technology nodes down to 0.11 µm microlithography are described, with Voltage capability in the range from 40 to 100 V. High Current, severe Energy dissipation, extended life time and temperature range are going to require robust solutions in terms of thermo-mechanical stress capabilities for the interconnections and the BEOL (Back End Of the Line). Concerning High Voltage usage the Hybrid/Electrical Vehicles are going to require High Voltage Gate Drivers Technologies (600 V) and specific solutions for Galvanic Isolations (6 kV).

European Photonics Industry Consortium (EPIC) in cooperation with TEMATYS has conducted a study on utilisation of photonics in automotive applications. This paper describes in brief the main conclusions of this study related in particular to automotive driving, ADAS, interior electronics and the green car.

Today’s Automotive industry is vastly aligned on the premises that current society’s evolution on ways to communicate, manage data, and interact with intelligent objects is going to define the future vehicle—life on board-services and the corresponding interior’s performance value. In fact, synchronized with the technological evolution, the way how population establishes relations between among intelligent objects (nomadic devices and home appliance, vehicle, etc.) is changing extremely fast towards extensive “self” concept and setting basis for new era of customer satisfaction requirements. Aging and Youngers car users of today are getting used, in the aside connected world of the car, to customized propositions, that allow multitasking, that provides them an Eco systemic functionality always available, symbiotic and interactive, a life companion (quantified self) that cares about. The Seat as part of this new Car interior is an essential enabler of such revolution as provides the unique opportunity to be one the closest vehicle part to the occupant and the first one in body contact surface, providing closed loop actions and information needed or supporting the-Life on Board-Services.

Technology doesn’t work very well. Car manufacturers face deepening challenges with technology as customer frustration with their multimedia systems builds. Technology issues are now the most prevalent type of problem with new vehicles.

Today, a successful automotive cockpit and HMI design is not only a beautiful and distinctive design, but also a design which reduces driver distraction and cognitive workload. Comfort and safety are at stake. Therefore, for OEMs, there is more and more value in performing quantitative measurements (reaction time, eye sight orientation, pupil diameter, etc.) of user-to-machine interactions. When done at an early development phase, these measurements are helpful in validating or rejecting HMI design concepts. ESG has started an internal project to help OEMs doing so. The project is to develop an HMI prototyping framework which includes all the facilities for bringing a potential user in front of a candidate HMI concept, and measuring how they interact with each other. The framework is able to host a wide variety of HMI prototypes, including multiple screens, innovative input methods and various actuators. After a test campaign is achieved, with several users, the framework enables statistical processing of measurement results, for later analysis (by ergonomists, for instance). ESG has finished the first phase of this project. They are ready to show a proof-of-concept demonstrator of this framework.

With the development of smaller and more cost effective electronics components, more and more applications in the consumer, but also in the automotive industry become possible. This has already led and will further lead to an increase of electronics content in vehicles and the linked infrastructure. The French industrial federation for electronics, electrics and communication, FIEEC, accompanies its members in this evolution. This paper describes in brief the interesting domains and the opportunities attached to them.