Self-driving cars: facing the realities and challenges

Also, download this story from the electronic issue here

Automotive makers are taking a shot at smart, handsfree cars that are forecast to ply the roads soon. However, are these geeky cars also ready to face the challenges on the road? Angelica Buan asks in this article.

A study by IHS Automotive predicts that total worldwide sales of self-driving cars (SDCs) will grow from nearly 230,000 in 2025 to almost 12 million in 2035. Of this, 7 million will have both driver control and autonomous control; and close to 12 million, only autonomous control.

In a nutshell, by 2035, there could be nearly 54 million SDCs dominating the highways. Give it a decade or two and nearly all vehicles in use are likely to be SDCs, or driverless commercial vehicles, said IHS in its report: Emerging Technologies: Autonomous Cars - Not If, But When.

The first batch of SDCs will have limited selfdriving, or the car will be allowed full control of all safety-critical functions under certain traffic and environmental conditions, including auto-pilot for highway travel and parking.

Nevertheless, some automotive makers are hoping to produce SDCs within the next five years. In fact, the technology of semi-automation is already becoming part of the in-car technology trend seen in smart vehicles bedecked with connectivity features. The semi-automation features help the market become comfortable with the concept of SDCs, until it is ready for fully automated cars, experts say.

Meet the new kings of the road

These turn-of-the-future cars are upstaging traditional cars. At the recent Consumer Electronics Show (CES) held early this year in Las Vegas, US, SDCs wowed the avid techie crowds.

German automotive maker Audi AG’s self-driving car technology was one of those showcased. The piloted driving concept, which utilises a combination of various sensors, relieves the driver of driving duties at over 110 km/hour. It can also initiate lane changes and passing manoeuvres, as well as accelerate and brake independently.

Providing safety features for the car, the sensors, many of which are close to production, include long range radar sensors of the adaptive cruise control (ACC) and the Audi side assist (ASA) that keep watch of the front and rear of the vehicle. Another two midrange radar sensors, at the front and rear respectively, are aimed to the right and left to complete the 360 degree view. Laser scanners are mounted within the Singleframe grille and the rear bumper skirt. These scanners deliver redundant information to provide detailed recognition of static and dynamic objects during piloted driving.

Mercedes Benz, German luxury vehicle manufacturer and a division of Daimler AG, also unveiled its F 015 autonomous driving car at the CES. Powered with two electric motors and designed for fuel cell technology, it can reach 125 mph, with sustained output of 163 mph. The four-seater car has swivelling seats, and interior concept with lighting effects and concave surfaces. It also provides an option for the driver to get behind the wheel.

Debuting at the CES was Volkswagen’s e-Golf featuring a self-parking capability known as Trained Parking. It utilises a camera, mounted in the base of the rear-view mirror of the e-Golf, to scan a frequently used path into a parking space. From then on, the parking process is executed semi-automatically by sensors and a computer. In the future, it will also be possible to semiautomatically park above an inductive charging station. In another evolutionary stage, the driver will be able to remotely monitor the operation with a smartphone.

Meanwhile, the North American International Auto Show, held in Detroit, US, saw Chinese car maker Guangzhou Automotive Group (GAC) displaying its self-driving plug-in car concept. Called Witstar, the autonomous car is powered by a battery and a gasoline engine. The car’s electric range is about 62 miles; and its range with fuel can expand up to 372 miles.

A novel highlight of the car is that it comes with a built-in fish tank in the rear seat. GAC came to the US’s biggest show this year, but officials didn’t indicate if they have any plans yet for the US market.

Likewise, introduced to the market is the Innovation Demonstrator (ID15) concept of US-based Johnson Controls, which features more than 30 technologies for autonomous vehicles. The ID15 features a rotating table that allows the driver to interact with second row passengers. Smart surface technologies are included with integrated control technologies enabling activation and control of various functions throughout the interior and serve as an alternative solution to conventional switches. Illuminated surfaces also differentiate the interior of the vehicle, and LED lighting across the instrument panel and along the side of the doors help improve the lighting ambience, with synchronised functionality to the Auto-drive mode.

Collaborative efforts

Researchers from Japanese car maker Nissan’s US Silicon Valley Research Centre and NASA (National Aeronautics and Space Administration)’s Ames Research Centre in California have formed a five-year R&D deal, targeting the first test by the end of the year.

The collaborating team will test a fleet of zero-emission autonomous vehicles to demonstrate proof-of-concept remote operation of autonomous vehicles for the transport of materials, goods, payloads and people.

The outcome of the partnership is said to benefit NASA’s development of its rovers and space vehicles. Likewise, it will also help Nissan’s development of autonomous drive technology that the company is aiming to introduce to consumers beginning 2016 up to 2020.

Also by 2020, US-headquartered Internet service provider Google is expecting to market SDCs, teaming up with major car manufacturers like General Motors, Ford, Toyota, Daimler and VW.

The Google prototype is simple, according to its developers, and is described as a practical solution for physically challenged people who are not able to drive.

Reality checks for SDCs

SDCs are the stuff that dream vehicles are made of, but the technology may not be perfect. The genius vehicles are bracing for challenges, according to industry observers.

Firstly, driverless vehicles come with a hefty price tag. A bulk of the price is the cost of the electronics technology, which will add between US$7,000 and US$10,000 to the price in 2025, according to the IHS study.

It also says that the figure will fall to around US$5,000 in 2030 and about US$3,000 in 2035, when no driver controls are available.

Nevertheless, automotive makers are also looking at ways to lower costs of SDCs, such as integrating cheaper yet efficient systems to function as currently used sensors and laser scanners.

Researchers from the University of Michigan (UM), led by Ryan Wolcott, and Ryan Eustice, have developed a new software system that may help forego the costly laser scanners used in SDCs.

The technology is said to enable cars to navigate using a single video camera that delivers the same degree of accuracy as costly laser scanners.

The new system converts the generated real-time map data into a 3-D picture much like a video game. The car’s navigation system can then compare these synthetic pictures with the real-world pictures streaming in from a conventional video camera.

The researchers have fine-tuned it to include a system that is able to process a massive amount of video data in real time. Trials have been undertaken on a system built out of graphics processing technology; and it managed to successfully provide accurate location information.

At the CES, an automotive computer was also launched by American technology firm Nvidia. Called Drive, the computing equivalent of a V12 for self-driving vehicles, is expected to usher autonomous vehicles onto the mass market.

The company offers two models: Drive PX for autopilot capabilities that adapts Nvidia’s mobile computing technology for autonomous vehicles; and Drive CX that has digital cockpit systems with the latest graphics processing architecture to deliver photorealistic graphics for applications including 3-D maps, landmarks and nextgeneration human-machine interfaces.

The Drive PX’s image processing technology will be used by the car to build a map of its surroundings, enabling it to auto-park, as well as detect other vehicles on the road by type, make and model, and to avoid obstacles or collisions.

Causes of concern for SDCs

While there are solutions offered for the high costs, a further worry is the lack of appropriate road regulations. Despite assurances from manufacturers that safety features are in place, automated cars are still a cause for concern for road mishaps.

In the UK, the government has started to pen rules to ensure that the driver/passenger in a SDC is alert, fit to drive, and capable of taking over the wheel at any time. Reports say that authorities are test driving these cars, in aid of the guidelines.

Meanwhile, there are the legal issues, such as the driver’s liability should the car be involved in an accident; and data privacy issues, considering that the GPS and other on-board connectivity devices store tracking information and personal information. Furthermore, concerns also centre on the flimsy protection from potential cyber threats.

Last but not the least is that these new breed of cars (such as the Google car) have yet to be proven reliable under certain weather and environment conditions.

One automotive maker that is looking at improving the above is Mercedes. It says that it is enhancing the data processing capabilities and sensor quality of its SDC model to ensure that the systems will work reliably under heavy downpour, severe winter conditions and in darkness.


Home | Terms & Conditions | Privacy Policy | Contact | Webmail | Site Map

Copyright (c) 2014 All rights reserved.