International Transportation (67) 2 | 2015 43 Automated driving - right-across the USA Driver assistance systems, Delphi long-distance drive, enabling technologies, radar, lidar sensors, V2X communication In spring 2015, automotive supplier Delphi Automotive, a manufacturer of sensors, automotive electrical components, electronic control systems, active driver assistance systems and engine technology, demonstrated the power of modern automotive engineering. After their successful driving debut at the CES in Las Vegas, engineers lined up what would be the longest automated drive to date. In March, a modified Audi SQ5 left San Francisco on an almost completely automated journey from the West Coast to the East Coast of the USA. Author: Thomas Aurich L oaded with Delphi technology, the- test vehicle covered around 5,000- km on the east-west drive, the longest journey ever completed by an automated vehicle in North America (figure 1). Sure, it was a demonstration trip, but it was also an opportunity to test and validate the systems and explore its possibilities by exposing the technology to real-life situations and circumstances that are hard to test in a closed experimental setup. During the journey, the engineers collected enormous amounts of data and established that the basic technologies for improving the active safety of cars are already available. The data obtained allow the design options to be expanded and improve the understanding of the complex technology. First test drives exceeded expectations Even before the transcontinental journey, Delphi had been using the vehicle in Mountain View, California and the surrounding area, as well as during the CES in Las Vegas. There, the car had already been confronted with a number of traffic conditions such as city crossroads, pedestrian crossings, unexpected obstacles, vehicles turning across its path, and more (figure 2). On top of that came a range of different weather and road conditions. The vehicle overcame these potential problems with flying colors. Technical equipment in the demo-vehicle The prototype autonomous driving car was equipped with a whole range of different sensors, using cutting-edge technologies and features, some of which are already available on the market. These included, for example, collision avoidance systems, integrated radar and camera systems, a rear-end collision warning device and lane departure warning equipment. These specific features will not all necessarily appear on future series-produced versions, but the move from active driver assistance technologies to automated driving brings with it daunting technical challenges in many different areas in parallel. The aim therefore was to All photos: Delphi Automated Driving SCIENCE & RESEARCH International Transportation (67) 2 | 2015 44 SCIENCE & RESEARCH Automated Driving integrate the systems into the car in the best possible way, taking into account any automotive engineering issues. Radar, camera and lidar sensors Radar, camera and lidar sensors (LiDAR: Light Detection And Ranging) detect changes in the environment around the vehicle. Radar: The vehicle uses a combination of six electronically scanning radar (ESR) units and four short-range radar (SRR) units. The ESRs specialize in long-distance sensor functions such as adaptive cruise control and cross-traffic detection. A forward-facing ESR is integrated behind a front license plate that is transparent to radar, while the front left and front right ESRs are located behind protective glass. The ESRs on the rear left and rear right are positioned near to the C-pillars, and one rear-facing ESR is integrated into the bumper. The SRRs are located behind cover plates at the four corners of the vehicle. Lidar: In contrast to rotating lidar units that protrude from the car, as used on many other independent platforms, vehicles working with Delphi technologies have six lidars that are integrated into the outer shell of the vehicle. This approach makes 360-degree coverage possible while maintaining the vehicle’s aesthetics. The lidars generate a high-resolution point cloud that contributes to general object recognition, particularly in busy urban environments. The forwardlooking lidar is integrated into the radiator grille, while the front left and front right lidars are located behind protective glass, and the rear left and rear right lidars are positioned near to the C-pillars and one lidar directed to the rear is integrated into the bumper. Each lidar is coupled with an ESR to effectively merge radar and lidar data. Sensor fusion: The perceptual system of Delphi’s automated vehicles improves efficiency by combining data from various sensors. As radar, visual and lidar sensors each have specific strengths and weaknesses, fusing the sensor data makes it possible to obtain a more accurate picture of the driving environment with a more robust detection of vehicles, pedestrians and other objects. Radio-based V2X communication: For automated driving, Delphi platforms use dedicated short-range communications (DSRC) to communicate with the infrastructure such as traffic light systems (V2I), with other vehicles (V2V) and even pedestrians (V2P). V2X communications provide redundancy, which is particularly useful in urban environments where there are a large number of traffic lights, vehicles and pedestrians. Intelligent software algorithms ensure that the systems always make the right decisions even in the most complex driving situations - for example when filtering into motorway traffic, crossing an intersection, or overtaking a cyclist. Multi-domain controller and Ethernet important components of-automated driving The automation of driving is one of the factors that cause the volume of data traffic in cars to soar. Increasing interconnectedness both inside and outside the vehicle requires new solutions for the architecture of electrical and electronic (E/ E) systems. The heart of tomorrow’s automated driving solutions will be the main driver assistance control unit, the multi-domain controller, which uses the latest high-performance processors. The multi-domain controller collects a whole range of sensor information from the vehicle’s environment for central analysis before the results made available to the onboard assistance systems. The controller is thus the pivotal interface for all functions in automated driving. The design and architecture of the control unit are part of Delphi’s innovative concept for the fast and secure processing of large volumes of data. Usually, the information is processed by various different control units that are all separate from each other. Ethernet is going to be an important standard for both the vehicle’s internal communications network and the connection systems. Thanks to its high bandwidth, it should eventually be able to replace bus protocols such as MOST and FlexRay, at least in part. Based on a study carried out in 2014 by ABI Research, an intelligence company focusing on the technology market, the use of network technologies in new vehicles will have risen still further by the end of this decade. A realistic scenario is that by 2024, the implementation of the Ethernet standard in new cars will take up almost 40 % of the connections. Different roads, different weather conditions The automated drive from the West Coast to the East Coast of the United States - from San Francisco to New York - was the first BACKGROUND INfOMATION The “Development Track” from Wuppertal to Silicon Valley The Delphi site in Silicon Valley was set up to-further accelerate the development of the portfolio that covers products concerning safety, environmental protection and interconnectivity. The Valley offers not only the flexibility to extend cooperation with an increasing number of technology partners, but also direct access to a rich pool of talent. The Technology Center in Wuppertal, Germany, is one of the largest in the Delphi organization, specializing in the development of electronics, E/ E architectures and system integration. Research and development activities in Germany and neighboring European countries ensure that technologies related to automated driving are being developed in line with the needs of European market. Figure 1: Route of the Delphi coast-to-coast automated drive Figure 2: Autonomous driving vehicle turning into traffic International Transportation (67) 2 | 2015 45 Automated Driving SCIENCE & RESEARCH such trip to be undertaken by an automatically driven vehicle. The Delphi engineers used the trip to collect and evaluate the data and information that will aid the development of technologies for active safety, which is the fastest growing area in the automotive industry. The team collected nearly three terabytes of data - the equivalent of about one-third the amount of data stored in the Library of Congress in Washington. The nine-day trip passed through nine US States and Washington D.C. On its route, the demonstration vehicle was exposed to diverse and challenging driving conditions: complex traffic situations, such as roundabouts, road works, bridges, tunnels, aggressive road users and different kinds of weather, but also to constantly changing road conditions, from bustling big cities to lonely cross-country roads (figure 3). The vehicle covered 99 % of the total distance of 3,400 mls (about 5,400 km) in fully automatic mode (figure-4). It showed that it could cope well with rain and snow, thanks to the combination of different systems that are active under different conditions. For example, on snowcovered roads, the radar takes over the task of the cameras, and V2V or V2X can detect broken-down vehicles even in conditions where the cameras only sees a field of white, and pass on the additional information to the vehicle systems. These products for active safety are already being used in vehicles and help steer the car safely through different traffic scenarios. The future To gain an idea of the different parameters involved in automated driving, it is useful to study a relevant guideline, such as the NHT- SA’s five stages of automated driving. • Level “zero” represents non-automated driving. • Levels 1 and 2 are already on the market and include technologies such as radar, optical systems, adaptive cruise control and lane departure warning. The driver is still fully responsible, including monitoring the environment and controlling the vehicle. • In Level 3, the driver may cede full control to the system in safety-critical situations and under certain traffic conditions. Although the driver can still intervene if necessary, he or she no longer has to pay attention to the road all the time. • Levels 4 and 5 correspond to high levels of automation, where the car intervenes on its own if the driver does not respond to the situation. The vehicle can perform all safety-critical driving functions automatically and “keep its eyes on the road” throughout the journey. The developers at Delphi are convinced that these innovative technologies have an important influence. Moreover, there are already a number of studies that attempt to describe the impact on society also in terms of numbers. There are four areas in which automated driving will improve our lives: safety, comfort, practicality and efficiency (labor productivity and fuel consumption). Safety is the area where to most significant improvements are to be expected. Customers are increasingly asking for active safety and would prefer not to have to constantly concentrate on the road situation while driving. This is why people will probably be prepared to pay a price for additional benefits such as these. Automated driving will appear in vehicles step-by-step, and manufacturers will be upgrading their vehicles accordingly, depending on available technology and market demand. Higher acceptance by consumers and suitable legal frameworks will certainly continue to drive this demand. ■ Rainer Denkelmann, Dr. Chief Engineer Global, Advanced Engineering, Body&Security, Delphi Deutschland GmbH, Wuppertal (DE) rainer.denkelmann@delphi.com Serge Lambermont, Ing. Physics Director Automated Driving, Delphi Electronics & Safety, Delphi Automotive, Silicon Valley (USA) serge.lambermont@delphi.com Thomas Aurich, Dipl. Oec., Chief Information Officer, Delphi Deutschland GmbH, Wuppertal (DE) thomas.aurich@delphi.com Figure 3: Autonomous driving vehicle in different situations on a highway Figure 4: Delphi car at the finish in New York