By Lee Hsin-Yin, CNA staff reporter

It may be 3 a.m., but 55-year-old Ou Hsiu-chu is wide awake, scribbling notes and taking photos of an orange and white vehicle the size of a mini-van easing its way down an exclusive bus lane on Xinyi Road in Taipei.

Ou is monitoring the first tests of a driverless bus carried out by Taipei authorities. The French-made vehicle only carries up to 12 passengers and operates at a tortoise-like 10 kilometers per hour, and Ou is not totally sold.

"I'm not sure about driverless buses for now because I haven't seen enough evidence that shows they are safe under less controlled circumstances," said Ou, the head of the ward where the tests were held, after taking a 5-minute test ride at the city's invitation.

The tests did demonstrate, however, the vehicle's ability to precisely detect environmental conditions on a Taipei street -- a breakthrough in Taiwan's push for smart vehicle development.

People in academia and the public and private sectors generally agree that this artificial intelligence-powered industry is worth pursuing because Taiwan is strong in information technology and has a well-established supply chain and local talent devoted to the field.

They believe that once demand is created under the central government's strategic guidance, Taiwan can become a major world player in driverless car technology, raising its profile as a knowledge-based economy to attract foreign investment, they said.

But the reality is more complicated.

DIFFERENT AGENDAS

Despite the general optimistic vision, reaction to the Taipei tests revealed friction between different stakeholders over driverless car goals and strategies that could derail Taiwan's development in the field.

For Taipei, the city sees autonomous vehicles as a way to improve the quality of life of its citizens but also as a source of competition on the innovation front with other major cities in Taiwan planning to run similar trials.

In the near-term, "the goal is to have all driverless bus-related technology and infrastructure tested to see how it fits our streets within one year," said Lee Wei-bin, director of the city's Department of Information Technology.

Longer term, the city hopes driverless buses will operate alongside regular buses to meet passenger demand at specific times, such as late at night or during rush hour, according to Taipei Deputy Mayor Lin Chin-rong.

CREATING A MARKET

While the city government is hoping to take the lead in developing an advanced transportation system, Martin Ting, the president of 7Starlake Co., the Taiwan representative of the French vehicle used in Taipei's test, is eying its market potential and creating opportunities for Taiwan to link up with the world's AI sector.

Ting said the most pressing issue following Taipei's bus test was to solicit greater resources for the industry so that manufacturers like 7Starlake do not miss out on a market that could determine Taiwan's economic future.

Taiwan is well-positioned in driverless car development, Ting argued, because it is already capable of producing key components of the vehicle, such as its IT panels, routers, motors, and battery management systems.

He was even optimistic that half of the components used in the vehicle made by France's EasyMile could be supplied and assembled in Taiwan within the next 12 to 18 months, creating a driverless car supply chain.

SUSTAINABLE DEVELOPMENT

Ting's vision is possible and practical, said Chen Cheng-Foo, former senior general manager of Hua-chuang Automobile Information Technical Center Co., which has been deeply involved in Taiwan's driverless car development.

But that business model is not going to benefit Taiwan much, Chen said, unless it can eventually develop its own research, development and service capabilities to keep from simply becoming an original equipment manufacturer for other countries.

"The reason to invest in driverless cars is to drive the industry," he said. "My advice is that the government takes the initiative to give a clear policy framework so that all the talent and resources can be integrated."

The government should also help define a niche market for the cars in order to make the industry sustainable, he contended.

For academia, the goal is to pursue Chen's suggestion and demonstrate Taiwan's ability to make its own smart vehicle, as personified by Li Kang, an assistant professor in NTU's Mechanical Engineering Department.

Li, a key member of NTU's driverless car development program that is partnering with Taiwan's Industrial Technology Research Institute (ITRI), said the school is trying to prove that the self-made driverless bus it is developing can be just as competitive as the EasyMile model.

NTU and other schools are also partnering with Ting and his company, but with seemingly different objectives. NTU thinks studying the French bus can help it with its own research, while Ting wants help in adjusting the parameters of the French vehicle to react better to the local environment.

REGULATORY RESTRAINTS

Beyond the discord in goals and strategies, almost all involved point to existing regulations as the biggest obstacle to the industry's development in Taiwan.

The main complaint is that prototype cars are currently prohibited from conducting tests in a more complicated environment to learn through the collection and analysis of data.

In response to the criticism, the Ministry of Transportation and Communications said the level of intelligence of autonomous vehicles remains unclear, which makes it difficult, for example, to draft standards for vehicle safety certification.

"The Xinyi Road test shows that this car can only operate extremely slowly (10 kph) in a highly-controlled environment," said Lee Chao-hsien, a specialist of the ministry's Department of Railways and Highways.

"This is a far cry from the traffic reality in Taipei, you know that," he said, stressing nonetheless that the ministry is open-minded on the issue and gathering information on how other countries regulate the driverless car industry.

NTU's Li doesn't buy the government's logic, saying that unless regulations are loosened to allow autonomous vehicles access to more complicated environments to learn by itself, significant progress in the development of AI-driven vehicles is impossible.

To facilitate dialogue among the central government, industry and academia on the subject, Kuomintang lawmaker Hsu Yu-jen, a long-time advocate of autonomous vehicle development in Taiwan, has vowed to establish a cross-party alliance by the end of 2017.

PUBLIC CONFIDENCE

But at the end of the day, Li believes public confidence in the system is key to whether it can take off in Taiwan, and he cautioned stakeholders not to over-promise what the cutting-edge technology can do.

"The tests we have seen recently are at a preliminary stage, and we need better infrastructure and step-by-step adjustments," he said.

And that is what Taipei intends to focus on next. It plans to apply for a specific frequency range from the National Communications Commission, which regulates the country's communications and information industry, and use that as a channel for traffic lights to send signals to the smart bus.

As a result, besides reading conditions around it through embedded cameras, the bus can receive confirmation of a stop signal from another source.

Through those efforts, Taipei is hoping to allay the concerns of people like Ou because they will enhance the vehicle's capabilities to react accurately to traffic lights, a variable not included in the first five-day test.

Yet even that test may have already contributed to some momentum in public support.

"The autonomous vehicle wasn't as terrifying as I had imagined," said 22-year-old Wang Tsung-wei after taking the test ride.

"I could tell that the bus was a bit confused when sensing something new on the road, but everything went just fine," Wang said. "I would very much like to try it again in the future."