AI Self Driving Technology

A Chinese VC’s Take on Investing in Autonomous Driving

Synced recently spoke with Delian Capital Senior Vice President Xuesong Fan about the current status of automated driving startups. Fan graduated from Harbin Engineering University and worked for years on Chinese satellite engineering. In 2015 came down to earth, applying his experience to self-driving car investments.

Just say the word “AI” and investors’ ears perk up, especially if the subject relates to a hot market like self-driving vehicles. Everyone wants a piece of the action in a tech that is transforming an industry.

Synced recently spoke with Delian Capital Senior Vice President Xuesong Fan about the current status of automated driving startups. Fan graduated from Harbin Engineering University and worked for years on Chinese satellite engineering. In 2015 came down to earth, applying his experience to self-driving car investments.

Fan and his team focus on automation, flexibility, information technology, and intelligence in the automotive industry. Over the past two years Delian Capital has funded CalmCar, which develops advanced driver assistance systems (ADAS) based on monocular cameras; and Abax Lidar Co, which focuses on chip-scale automotive LiDAR solutions.

Execution, Decision Making, Perception

Many technologies and industry trends develop fairly gradually, allowing players to adopt, whereas self-driving is disruptive tech expected to completely remake the automotive industry. We’ll examine the tech’s investment opportunities from three angles: execution, decision-making, and perception.

Execution includes fundamentals such as powertrain, braking, steering and security systems — technologies with almost a century of history. From an investor’s perspective, there is little opportunity here, unless they encounter a team with more than ten years experience in automobile manufacturing with its own technical attributes and industrial capabilities. Fan’s company gave up on investment in execution and decided to choose between perception and decision-making.

Self-driving car technologies are being developed by Internet companies, e.g., Google and Baidu; Industry, including automotive manufacturers Ford, Mercedes-Benz and BMW; and in Specific Scenarios, such as low-speed patrol cars or sanitation trucks.

Regarding autonomous driving decision-making, abundant resources, funds and strong tech are required for startups to solve complicated technical problems and compete with automotive giants. However, when Fan’s team was making its investment plans, the decision-making process for autonomous vehicles was not particularly mature in theory or practice. The team decided it was too early to fund start-up companies in this area, and turned to perception.

Perception is the process of obtaining physical information from the surrounding environment using various sensors, which is not only essential for autonomous vehicles but also required in other smart device scenarios such as robots and drones. Examples of perception solutions include cameras, LiDAR, millimeter-wave radars, infrared sensors, etc.

Seeking Camera Companies

Among perception solutions, cameras are the only sensors with recognition capability that can be optimized using iterative algorithms. Visual recognition is crucial in self-driving for detecting pedestrians and obstacles. In terms of camera selection, the monocular camera is top-ranked by Delian Capital because it has good recognition capability, and unlike binocular cameras it is less affected by environmental conditions such as rain or extreme sunlight.

Once the camera is selected, the visual recognition algorithm should be determined. For traditional pattern recognition, Israeli-based Mobileye has been a global leader since 1999, and has collected a massive amount of data and iterations from driving scenarios around the world. As it is difficult for a company to quickly become competitive in this area, Fan’s team decided to focus on companies that apply deep learning for pattern recognition.

As AI has developed over the past years, deep learning is now playing an increasing role in self-driving car scenarios: Problems can be solved through model training processes, where both high-quality training models and data are keys to achieving good results. Moreover, Fan says he pays close attention to the automotive industry background of a company’s founding team, since this is important for understanding industry rules and engineering properties in the field and maximizing team productivity.

Fan discovered CalmCar in 2016, and believed the company matched his team’s investment expectations: Company CTO Xiaojun Xie is a PhD in mathematics and has years of experience in the development of deep-learning algorithms in Silicon Valley; The company also cooperates closely with automotive agencies and OEMs in China for data acquisition.

Finding Opportunity in LiDAR Company

At the end of 2016, Delian Capital completed an overall arrangement for funding ADAS, and started exploring opportunities for LiDAR perception sensor chips. This quest was satisfied when a new LiDAR type with a silicon photonics platform for analog-to-digital conversion emerged in the second half 2017. Abax Lidar’s focus on such LiDAR sensor chip development and the company’s background and technology roadmap met Fan’s team’s investment criteria.

Even though the chip had not yet been developed, Fan and his partners had confidence in Abax Lidar’s founders and core technicians, considering their experience in LiDAR chip development and production. Delian Capital was pleased to bet on the company, and whether the project is ultimately successful or not, it provides valuable training opportunities for potential Tech talents, not only in the self-driving car industry, but also across other fields.


Localization: Tingting Cao | Editor: Michael Sarazen

0 comments on “A Chinese VC’s Take on Investing in Autonomous Driving

Leave a Reply

Your email address will not be published.

%d bloggers like this: