Taiwan Gearing Up for Driverless Cars


Fully automatic buses are already being tested in Taipei for late-night service.

Backed by its solid base in ICT (information-communications technology), Taiwan is gearing up for the coming trend of driverless cars, which is expected to revolutionize transportation worldwide within 10 years.

Taiwanese automaker Yulon Motor Co. has been among the leaders domestically in this effort, working through its subsidiary, the Hua-chuang Automobile Information Technical Center, the unit responsible for developing Yulon’s indigenous Luxgen-brand models. In August this year, the company showed off a driverless car on stage during Luxgen 2017, an annual technology forum held by the company.

The prototype, a Luxgen S3 EV+ electric car, was the fruit of an R&D alliance led by Hua-chuang and consisting of over 100 auto components and parts firms. The car is furnished with ADAS (advanced driving assistance system), anti-crash and auto-parking functions, and internet access, the so-called Internet of Vehicles.

“Cooperation between Taiwan’s ICT and auto industries can give birth to an industry with great potential, in addition to helping the partners certified by Hua-chuang to tap into international markets,” says Yulon Vice-CEO Chen Kuo-jung.

At the Autronics Taipei exhibition in April, the government-sponsored Automotive Research & Testing Center (ARTC) unveiled its self-developed auto-driving technology, installed in an electric golf cart and capable of auto cruising, lane change, auto parking, and automatic car retrieval. Furnished with one 3D LiDAR (radar with laser technology) and camera on the front end of the car roof, as well as four ultrasonic sensors, the vehicle is capable of braking upon detecting pedestrians, returning to the right-hand lane in case of aberration, and changing lanes to avoid obstacles in front.

Driveless buses are now being tested for late-night service along Taipei’s XinYi Road. PHOTO: TAIPEI DEPT. OF INFORMATION TECHNOLOGY

Major Taiwanese ICT firms are scrambling to join the supply chain for driverless-car production, viewing it as the potential next growth driver for the global ICT industry.

The integrated circuit (IC) industry is a good example. MediaTek, Taiwan’s leading fabless IC firm, has announced plans to develop chips for use in ADAS, high-precision millimeter-wave radar systems, auto information and entertainment systems, and auto information-communications systems.

Hsieh Ching-jiang, MediaTek’s vice chairman and president, notes that the company will offer an integrated solution, sparing automakers the time and cost of integrating products and systems from different suppliers firms. The first batch of auto chips has been sent to prospective customers for certification, a process that may take a year or more.

MediaTek has followed in the footsteps of some of the world’s leading chipmakers, including Qualcomm, Intel, Nvidia, and Micron, in seeking to meet the growing demand for automotive chips – a demand that is expected to increase sharply with the advent of driverless cars. Micron has estimated that a driverless car will need 200 gigabytes (GB) of DRAM storage capacity, compared to the 60GB for current conventional models.

Taiwanese firms are also eyeing the huge potential demand for such components as auto lenses, which are essential for smart-car ADAS for applications in around-view monitoring systems, lane-departure warning systems, and head-up display. Cheng Sheng-chuan, chairman of auto-lens maker Newmax Technology, notes that most new car models in 2018 will be equipped with a semi-auto driving function requiring 36 lenses. Calin Technology, a pioneering auto-lens supplier in Taiwan, supplies major automakers in Japan, Europe, the United States, and China, with shipments expected to top 2 million lenses in 2017.

In addition, several Taiwanese communications devices manufacturers, including Wistron NeWeb, Cub Elecparts, Alpha Networks, and TungThih, have developed millimeter-wave radars for application in smart cars. Several chipmakers are also developing GaAs epitaxy chips for use in auto radars, as such chips feature high frequency, low noise, high efficiency, and low power consumption.

ADAS is considered to be the brain of driverless cars, integrating data gathered by various in-car sensing devices, including image sensors, radars, LiDAR, and GPS, to produce a 360-degree view of the vehicle’s surroundings. That “sensor fusion” is then transmitted to the in-car CPU for high-speed computing and analysis leading to controls on the car’s speed and direction.

The testing team with the driverless bus being prepared to serve Taipei’s night owls. PHOTO: TAIPEI DEPT. OF INFORMATION TECHNOLOGY

Driverless mini-buses

Meanwhile, the Taiwan government has started to prepare for the advent of the driverless car era, especially in mass transportation. On August 1, a French-made six-seat driverless bus, dubbed “EZ 10,” started running along the dedicated bus lanes of Taipei’s XinYi Road between DunHua and FuXing South Roads, after midnight on a trial basis. The bus has a cruising speed of 20 kilometers/hour, and is furnished with six LiDARs, capable of detecting all objects within a 50-meter radius. It can brake in milliseconds, without any blind spot, and has high-precision GPS with error tolerance of just one to three centimeters, compared with over one meter for ordinary GPS, enabling the bus to pull over at bus stops precisely.

Li Chen-yu, director of the Taipei Smart City Project Management Office under the city government’s Department of Information Technology, says the ultimate objective is to put driverless buses into service during off-peak hours, especially in the middle of the night. “That’s the time when driverless buses are especially suitable,” says Li.  “Humans are more likely to commit errors at that hour due to fatigue or poor visibility.”

By the end of this year, the city government plans to open an enclosed testing ground for use by driverless-car developers at the site of the projected Shilin Technology Park, which is still under construction. “We will offer a V2I (vehicle to infrastructure) environment, as many parties, both domestic and foreign, have expressed strong interest,” says Li. “ITRI (the Industrial Technology Research Institute), for instance, plans to test its automatic-driving software at this facility.”

Driverless cars have also been included, under the rubric of the Internet of Vehicles, in the “Plan for the Development of a Smart Transportation System” drafted by the Ministry of Transportation and Communications. Funded by a budget of several billion NT dollars over a four-year span (2017-2020), the plan calls for the application of ICT to reduce traffic congestion by 25% and the auto/motorcycle accident rate by 20%, while increasing the use of mass transportation by 10%. The plan envisions the creation of NT$30 billion (US$1 billion) in annual production value in related industries.

Although virtually every major hi-tech firm and automaker worldwide has joined the fray in the effort to develop driverless cars, the frontrunners are considered to be Tesla and Google.

Since the debut of Tesla’s “Autopilot” in October 2015, its electric cars with semi-automatic driving function have accumulated 200 million kilometers of driving distance. After receiving an injection of US$1.8 billion in investment from Tencent, China’s internet giant, in March, the company plans to accelerate the development of cars with a fully automatic driving function, the “level five” of automatic driving.

The fleet of driverless cars of Waymo, a subsidiary of Alphabet, parent company of Google, has been driving on the roads of California for years, accumulating a total driving distance of 3 million miles.

Many major automakers have rolled out models with semi-automatic driving functions, such as automatic safety-distance maintenance. They include Audi, General Motors, Ford, Nissan, Honda, Toyota, Daimler (including its Mercedes-Benz division), and BMW.

Multiple hurdles, however, remain to be overcome before driverless cars can become mainstream vehicles on the roads. First is the laborious and formidable task of collecting and processing massive data on local road and transportation conditions – data that needs to be constantly updated – for input into in-car computers.

The most difficult task is expected to be overcoming skepticism among members of the public regarding the safety of driverless cars, despite advocates’ claims that they are 100 times safer than human-driven cars by preventing accidents caused by driver fatigue, distraction, intoxication, or dangerous driving habits.

Still, the auto industry seems convinced that driverless vehicles are the wave of the future. Taiwan therefore needs to position itself well to take advantage of the benefits that trend may bring, both in terms of business opportunities for domestic industry and improvement in transportation conditions on the island.