A Collision-Free Automotive Society Through Autonomous Driving

Ms. Fang Fang


Fang Fang graduated in mechanical engineering from Dalian Nationalities University in China in 2008. She completed her master’s degree in mechanical engineering at the University of Tokyo’s School of Engineering in 2012. The same year she joined the Nissan Research Center (NRC) at Nissan Motor Co., Ltd., where she is engaged in the research and development of surrounding-environment recognition for autonomous driving.

Reasons for Choosing Nissan Research Center?

In Pursuit of an Ideal: A Collision-Free Society

It was Fang Fang’s desire to work overseas that first brought her from China’s Heilongjiang Province to Japan. She joined Nissan after graduate school, familiar with the company as one admired even in China for its success as a global automaker. Her decision was also influenced by a site visit to Nissan that she participated in after arriving in Japan, where she was impressed by the positive atmosphere in the workplace and the steps taken to ensure a woman-friendly work environment.

Naturally, joining the company was predicated on the opportunity to apply her graduate research. Fang’s strong interest in technology enabling a safe, collision-free automotive society led her to focus on researching the relationship between controls and driving errors in older drivers. Ultimately she hoped to realize the ideal of eliminating traffic accidents entirely.

Current Projects in a Nutshell

Developing Onboard Systems to Rival the Human Eye and Brain

Fang has been with Nissan for five years now. In her first year, after completing a variety of training courses, she was assigned to the field of automatic driving control. Here she was involved with projects like path-planning simulations to be used by autonomous vehicles when parking. Starting in her second year, she was put in charge of R&D of surrounding-environment recognition for autonomous driving. In fully autonomous driving, the three elements of driving—cognition, decision and actuation—are performed by the car instead of the driver. Fang’s research centered on cognition to correctly grasp conditions around the vehicle. This involves obtaining detailed information about the car’s surroundings through a combination of sensor components, including cameras, millimeter-wave radar and laser scanners. This data must be rapidly and correctly integrated to detect newly emerging risks, perhaps more quickly than even a human driver can do it.

“The goal is to recognize the vehicle’s surroundings in the same way as the human eye and brain and then process that information in order to make the correct decision,” Fang explains. “However, I feel that there are still several barriers that remain to be overcome before we reach that point.”

A Record of Successful Research

Advancing Step-by-Step, One System at a Time

Many challenges remain before the final objective is reached, but research is advancing step-by-step. At the end of 2015, following experiments with a real car in California’s Silicon Valley, recognition accuracy was successfully improved through fusion of multiple sensors’ outputs, among other techniques. One example of success in this area is developing the ability to detect objects after they disappear behind other objects and become undetectable to sensors, then re-establishing contact when they reappear.

“If pedestrians or other moving bodies enter a spot not visible from the vehicle and become undetectable, serious accidents can result,” Fang goes on. “The ability to sustain recognition of the presence of pedestrians, even if they can’t be seen, and then quickly distinguish them and detect possible risks once they reemerge is crucial. This comes easily to humans, but autonomous cars have difficulty with many aspects of the task. Through revisions to our existing logic and other improvements, we’re slowly but surely expanding the range of what cars can do.”

Incremental progress is achieved by gradually eliminating these spots where the sensors are constrained and realizing performance improvements in order to respond to the countless situations the driver of an autonomous car might encounter.

Current Challenges

Accurately Predicting the Actions of Other Road Users

Fang is currently focused on R&D aimed at correctly predicting the behavior of objects once they have been recognized by sensors.

“Humans look at the traffic around them and immediately assess which vehicles are trying to change lanes, which are about to slow down or turn left or right and which are about to stop,” she says. “In order to make this ability, which humans perform so smoothly, a reality in autonomous vehicles as well, our next goal is to incorporate movement patterns into the AI and develop it to the point where highly accurate predictions can be made based on the mutual influence between objects, among other things.”

Even if sensors become more advanced and the range of detection around cars expands, allowing highly accurate cognition, creating systems that can make appropriate decisions based on this at the same level as a human awaits at the next phase of development. No sooner is one challenge overcome than another appears. But Fang remains positive: “Thinking about the best way to deal with challenges, not to mention experimentally testing one’s ideas, is always interesting.”

Dreams for the Future

Helping to Protect People Through World-Leading Research and Technology

The world is abuzz with news about autonomous driving. Fang finds it exciting to know that so many people across multiple industries are dedicated to realizing the same dream. “The pride that comes from knowing that the technological development program I’m involved with attracts attention from all over the world is an excellent research motivator,” she says.

If fully autonomous driving becomes a reality, Fang believes, it may become possible to reduce the incidence of accidents caused by human error. It may also help lessen the burden on the environment with more efficient traffic flow brought about, for instance, by avoiding the sort of traffic congestion that arises when human drivers allow their vehicles to lose speed when traveling up a gentle slope. To make this dream for the future a reality, however, predicting and allowing for circumstances other than the ideal is also necessary.

“Today our R&D assumes that drivers obey the laws of the road,” Fang observes. “Going forward, though, we’ll need to think about how to deal with situations where the surrounding cars fail to observe traffic lights or signals. A long road lies before us, but I look forward to anticipating and meeting each challenge as it comes. ”

<A Note to Prospective Researchers>

“To create the car of the future,” Fang declares, “I feel that cooperation will be vital not only between people of different national backgrounds and genders but also between people with different cultures, lifestyles and backgrounds. Nissan’s corporate culture based on cross-cultural understanding and diversity encourages all sorts of people to exercise their full potential.”

Based on an interview carried out in June 2016.