ADAS to AVs - more progress in 2020, but with added reality check

Whether governments and politicians have the resources or the political will required to make such investment, and see this market develop in the short term, remains to be seen. However, there are clear signs that all interested parties are testing the water and trials are being run to test the technology and systems required. Indeed, Google has been one of the first to run trials of its technology on a self-driving vehicle in California, USA. Many of the vehicle manufacturers are running similar trials to a greater or lesser extent. 

Advanced Driver Assistance Systems (ADAS) that are here today point some of the way and there are cars on the market already that allow limited autonomous control in certain situations (like parallel parking, or traffic jam assist). And ADAS systems already employ radar scanners and cameras – technologies that herald a shift from traditional driver assistance toward totally autonomous control. As driving habits change and technology develops, human drivers will ultimately be replaced by safer and more reliable autonomous vehicles (AVs) - but look beyond 2030 for fully autonomous vehicles to be genuinely market ready. 

Safety is a key motivator of this technology. A number of manufacturers are looking to reduce the frequency and severity of accidents by utilising active and passive driving-assistance systems. These systems allow the vehicle to perceive its surroundings, interpret obstacles and critical situations, and assist the driver in performing manoeuvres. The aim is, ideally, to completely prevent accidents or at least minimise the consequences for those involved. 

Expect to see more rollout of ADAS systems in 2020, with companies also very active in driverless/autonomous concepts and fleet testing. 

Driver assistance systems use a combination of warnings and some degree of active intervention to help steer the driver away from trouble. Although the accent is on giving assistance to the driver rather than take control away, motorists are still wary about cars that supposedly drive themselves. While active intervention clearly holds many possibilities, it is also fraught with difficulty. 

The most common suite of driver assistance technologies available today includes adaptive cruise control, lane departure warning systems, and parking assistance systems. These are now 'cascading down' from premium segments to mass-market ones. Radar-based safety technologies such as advance collision warning and blind-spot detection are becoming common place as optional equipment on new vehicles.  

Driver assistance systems – either on the road or still on the drawing board – divide into three distinct categories: 

• Collision-warning systems – this is the original term for forward and side radar systems which alert the driver but does not control the engine speed. Driver support systems without active intervention can be viewed as a pre-stage to vehicle guidance. They only warn the driver or suggest a driving manoeuvre. 
• Collision-mitigation systems – in addition to sending out a warning to the driver (either through audio, visual or vibrating the steering wheel), these systems aim to assess the danger ahead and activate various active safety features, such as pre-tensioning the seatbelts. 
• Collision-avoidance systems – using these systems mean that some degree of control is taken over from the driver if he doesn’t react in time to avoid a crash. Although there are some technologies available today that fit inside this group. 

All of these driver assistance systems – from collision warning through to avoidance technologies – are feasible either now or at some point through this decade. However, there are significant problems in software control, not least how to make the various systems work together to form a true protection zone around the vehicle and then develop algorithms to determine if or when the system should intervene with the driving process. 

Collision avoidance and autonomous emergency braking (AEB) systems are the most popular advanced driver assistance systems (ADAS) among UK fleet managers, according to UK leasing company Arval.  

The updated (Mark 7) Volkswagen Golf provides a good example of just how far we have travelled in terms of advanced driver assistance systems (ADAS) fitment on mid-segment cars. The latest incarnation includes adaptive cruise control, a driver alert system, predictive pedestrian protection, lane assist and traffic jam assist. The latter enables semi-autonomous driving in traffic jams. 

Another notable ADAS feature fitted to the Golf is Emergency Assist. This notices if the driver is incapacitated and initiates various measures to rouse them in escalating stages culminating, if the driver remains inactive, in carrying out an emergency stop. 

Other optional driver assistance features on the updated Golf include Trailer Assist with Park Assist. Linked with the rearview camera, Trailer Assist technology helps the driver reverse with a trailer or caravan attached.  The system helps reverse a trailer or caravan into a parking bay or entrance from the street. To operate, the driver selects reverse gear and presses a button. The driving angle and subsequent possible driving angles are displayed on the instrument cluster. All the driver then has to do is operate the accelerator pedal and brake, while the system automatically controls the direction of the car, using Park Assist to move the trailer backwards into the required position. 

In its third iteration, Park Assist makes it possible to park semi-autonomously in any parking space that is parallel or perpendicular to the road; in addition, the system can exit from parallel parking spaces. For perpendicular parking spaces, not only is parking in reverse supported for the first time in a Golf, it is also now possible to drive into a parking space semi-autonomously. The first generation of Park Assist helped the driver by performing automated steering for parallel parking. The second generation of Park Assist added automatic steering for reverse parking into spaces perpendicular to the carriageway, and it could handle smaller parallel parking spaces. In addition, automatic exiting from parallel parking spaces was also possible. The third generation of Park Assist also makes it possible to park forwards into parking spaces semi-autonomously. 

Material-wise, we are seeing more centre console screens made of flat and curved glass. Although the use of glass for touchscreens creates a classy look, this illusion is soon shattered by smudgy fingerprints. The HMI (human machine interface) therefore needs greater refinement before glass screens become widespread. Glass screens do not currently provide the user with haptic feedback, causing driver distraction when double-checking activation of functions.

Another reason tomorrow’s car interiors could see more screens is the gradual disappearance of the exterior rearview mirror. Given that wing mirrors add weight, cost and wind resistance (at high speeds) to the car, it is not surprising that some automakers wish to eliminate them thereby creating a sleeker look. As suppliers anticipate a change to mirror regulations, many are developing mirrorless systems using vehicle mounted cameras and dashboard monitors.

Mirrorless cars use video systems and dashboard screens to increase the driver’s field of view, typically eliminating the blind spot by stitching together a panoramic view of what is behind and to the side of the car. Although mirrorless cars will take time to get used to, most solutions we have seen feature high-resolution screens located close to where a driver would glance to check the wing mirror. Some mirrorless technologies assist the driver further by automatically adjusting to reduce sunlight glare or intensifying levels of brightness while parking at night.

With future trends pointing to the elimination of wing mirrors and the addition of screens, such black rectangles can dominate the look and feel of a car’s interior. It has even been disputed as appearing a little dated; the opposite of a luxury interior designer’s objective. To some extent, flexible OLED (Organic Light-Emitting Diode) displays that blend into the interior can help (see below). A balance therefore needs to be struck. Greater use of projecting driver information onto head-up displays (HUDs) and augmented reality windscreens would further reduce dashboard screens.

Many disparate strands – government, infrastructure, technology, cost to consumer – have to come together to make autonomous driving a widespread reality. Regardless of claims by automakers, it will be many years before truly driverless cars become a reality and they can swivel their chair to face the back seat passengers. Although much of the technology needed to operate self-driving cars has been developed, the laws that allow such vehicles on our roads are some way behind. While some automakers are predicting fully autonomous cars by 2030, there is a still a lot of red tape to wade through before then. 

Meanwhile, AV road testing continues apace and get ready for more in 2020. Last year, Daimler received the authorities' approval for highly automated driving (SAE Level 4) test area road use in Beijing following extensive closed-course testing. It says it is the first international automaker to receive a road test license for L4 in the city. ZF says it is looking to develop platooning concepts as it moves to increase its operations with autonomous and electric driving for vans and commercial vehicles.  

The autonomous movement is also sparking a wave of alliances and M&A activity as industry participants move to ensure they have the rights assets and intellectual property to address the changing automotive value chain. Collaborations are spreading. For example, Apollo, Baidu's open autonomous driving platform, provides a solution that supports major features and functions of an automated vehicle. Daimler is a member of the Apollo Committee, subsequently broadening its collaboration to explore new fields in-vehicle connectivity services. Other signed-up members of Apollo include BMW, Analog Devices, Mobileye and Valeo. 

Waymo - formerly the Google self-driving car project - is piloting autonomous cars in Florida, expanding from its initial test cities of Novi, Kirkland, Washington, San Francisco and Phoenix. Since 2017, Waymo has been trialling a limited autonomous robotaxi service in Phoenix, Arizona with the intention of escalating into a large-scale robotaxi fleet. Waymo has already acquired a fleet of Chrysler Pacificas and Jaguar I-Paces to realise its ambition. 

Waymo cars are manually driven by trained operators, giving the team an opportunity to collect real-world driving data in heavy rain. When the cars venture beyond the closed track, they will mainly drive on highways between Orlando, Tampa, Fort Myers and Miami. 

Waymo has launched Waymo One, its commercial driverless taxi fleet of over 600 cars with safety drivers behind the wheel and the company says the fleet has grown to serve 1,000 riders in that time. Separately, Waymo recently revealed its cars have driven 10bn autonomous miles in simulation and 10m real-world autonomous miles in 25 cities. 

Waymo has also announced it will dedicate a factory in southeast Michigan to the production of level 4 autonomous cars — that is cars capable of driving without human supervision in most conditions — the company said it had settled on a location in Detroit. Separately, Waymo partnered with Lyft to deploy 10 of its vehicles on the ride-hailing platform in Phoenix. 

In Detroit, it tests driverless Chrysler Pacifica hybrid minivans produced in Windsor, Canada and shipped to Novi, where they are outfitted with hardware and software by Waymo and Chrysler engineers. 

Waymo also announced last year it would add up to 62,000 minivans to its fleet and said it had signed a deal with Jaguar Land Rover to equip 20,000 of the automaker's Jaguar I-Pace electric SUVs with its autonomous system by 2020. 

Baidu has debuted the Apollo Robotaxi in Changsha, Hunan province, with the first batch of 45 self-driving taxis officially starting trial operations on urban roads. Apollo Robotaxi deploys L4 Hongqi EV vehicles jointly developed with Chinese carmaker FAW Group. Ride hailing is now open for trial by the general public in Changsha, with the fleet of robo-taxis able to perform a range of intelligent self-driving functions, including changing lanes based on road and traffic conditions, assessing the movement of nearby vehicles and employing automatic avoidance when encountering aggressive overtaking by other vehicles. With safety as a top priority, a human operator is also assigned to each robotaxi to take control of the autonomous system in situations that could affect the safety of passengers or the vehicle itself. 

A fleet of 30 vehicles equipped with Baidu's Apollo autonomous driving technology has also begun trials in the city of Cangzhou in China's North China's Hebei province. The vehicles have been tested on 114 km of designated roads in the Cangzhou Economic and Development Zone, as part of the region's efforts to speed up development of emerging technologies such as artificial intelligence and intelligent vehicles and transportation systems. 

Chinese smart mobility company WeRide has begun large scale autonomous taxi tests in the city of Guangzhou in China's southern Guangdong province. The company, backed by the Renault-Nissan-Mitsubishi alliance, installed its latest Level-4 autonomous driving technology in a fleet of dozens of battery-powered Nissan passenger cars for trial in a 144 sq km area of the city.  The trial is being conducted in collaboration with local taxi company Guangzhou Baiyun Taxi Group, after the two companies reached an agreement in August. The WeRide vehicles are the same in appearance as the regular taxis operated by its partner and even have the same meters. As part of the trials, passengers can order a taxi via the WeRideGo app, with dozens of rides having taken place on the first day. 

Autonomous vehicle trials have been stepped up on dedicated roads and test centres across China this year, as the country looks to become a global leader in autonomous and connected car technology. 

Meanwhile in Russia, Google competitor Yandex is planning to increase its fleet of self-driving vehicles ten-fold to more than 1,000 cars. The internet services company is considering expanding its fleet of self-driving cars to up to 1,000 within the next two years in order to speed up tests on the new technology. Currently it has around 100 self-driving cars available for testing and is testing on Moscow roads. 

In Moscow, where drivers frequently flout road regulations, Yandex says that the cars have to learn how to navigate safely through chaotic road conditions and speeding drivers.  

A new test centre dedicated to autonomous and connected vehicles has opened in Xiangyang in China's Hubei province. The CNY1.6bn (US$228m) centre covers an area of 73 hectares and is understood to be the first of its kind in the country. The dedicated facility allows automakers to test leading connected and autonomous driving technologies in a broad range of scenarios, using 5G systems, precise positioning and real-time sensing. 

China is looking to become a global leader in autonomous and connected car technology and trials are being stepped up on dedicated roads and test centres across the country. 

Share this article