Exclusive interview with DeepMap COO Luo Wei: In the era of autonomous driving, how can startups break through in the field of high-precision maps?

Recently, Silicon Valley high-precision map service provider DeepMap announced the completion of a $25 million Series A financing round, led by Accel, with angel investors GSR Ventures and Andreessen Horowitz following suit. It is reported that GSR Ventures and Andreessen Horowitz had previously invested $7 million in the company. So far, DeepMap has completed $32 million in financing.

36Kr learned that the DeepMap team members include professional engineers and product managers from Google Maps, Google Earth, Apple Maps, and related cloud services, LiDAR and other related fields. DeepMap provides high-precision maps, real-time positioning software and related necessary cloud services required for autonomous driving. Currently, DeepMap is working with some manufacturers in the automotive industry chain to help them realize safe autonomous driving technology faster.

If the various sensors of driverless cars are their eyes and ears, high-precision maps are their memory and teachers. The premise for driverless cars to be safe on the road is to be able to accurately understand the surrounding environment and identify it in real time. Among all environmental information, maps are the most important part. Without accurate and efficient high-precision map services, driverless cars cannot be safe on the road. Traditional methods of collecting, drawing and maintaining high-precision maps require very high costs.

This is where the opportunity lies for technology startups like DeepMap.

DeepMap COO Luo Wei told 36Kr that DeepMap has launched a new solution for the high-precision maps required by self-driving cars, which can quickly and cost-effectively provide one-stop services from map creation, map consumption to related cloud infrastructure. Specifically, DeepMap will focus on three directions:

• DeepMap will provide 3D centimeter-level high-precision maps;

• DeepMap will help self-driving cars "consume" high-precision maps;

• DeepMap will provide the necessary cloud service infrastructure.

As we all know, map providers such as Here and TomTom have been occupying their respective markets. Facing these giants, what is DeepMap's core competitiveness?

Luo Wei told 36Kr that traditional map vendors such as Here Maps still maintain traditional business models in the era of autonomous driving. In other words, they only invest in the first item of "providing high-precision maps" mentioned above. Helping autonomous driving cars "consume" high-precision maps and the necessary cloud service infrastructure is the most complex and laborious pioneering work in the field of high-precision maps in the future. DeepMap will integrate the perception data of various sensors including lidar and cameras, and convert the collected data into high-precision maps that can be "understood" by cars. DeepMap will open a set of API interfaces to car company partners to help partners' autonomous driving cars better "consume" high-precision maps. Compared with competing products, vehicles in the DeepMap platform can share information with each other, and maps can be updated in real time. At the same time, the data transmitted by the car to the cloud will be efficiently processed and fed back to the cloud and driverless cars in real time. This technology makes it possible for high-precision maps to be applied to driverless cars on a large scale and at low cost.

"The social significance of driverless cars is quite significant," said Wu Xiaqing, CEO of DeepMap. "But to enable driverless cars to run unhindered in the real world, accurate mapping and positioning technology is a necessary prerequisite, and it must ensure that they can operate safely and efficiently. Solving this problem takes a lot of time and money. DeepMap works with the automotive industry to achieve their autonomous driving R&D goals faster and more efficiently, and helps them reduce risks and save costs."

Martin Casado, partner at Andreessen Horowitz, said: "Self-driving cars will bring about a revolutionary change in our future. To achieve autonomous driving, high-definition maps, real-time positioning software and related necessary cloud services are indispensable. I am very excited to work with the DeepMap team. They have many years of experience in mapping technology, product vision, and massive big data cloud services."

In addition to the big-name investors, as CEO Wu Xiaqing mentioned, DeepMap actually adopted a crowdsourcing approach in drawing high-precision maps, which is also considered to be the industry consensus for developing high-precision maps. Luo Wei told us that several large automakers and Tier 1 suppliers have shown great interest in DeepMap's technology, and related cooperation is also under negotiation.

In DeepMap's view, in order to provide high-quality, high-precision maps and positioning guides for self-driving cars, the team needs to have a variety of technical strengths, including surveying and mapping, sensor fusion, machine learning, computer vision, SLAM (simultaneous localization and mapping), geometric modeling, and cloud architecture construction, etc. The DeepMap team brings together world-class talents in the above-mentioned technical fields, which is also the competitiveness of DeepMap that Luo Wei has always emphasized. After this round of financing, DeepMap will further expand and optimize its high-precision mapping and positioning technology platform, and continue to absorb talents in the fields of map surveying and autonomous driving to expand the team size.

From Horizon Robotics in China to Argo.ai and Aurora abroad, it has become a new trend for technology executives to leave and found startups. With core teams from Apple, Google, and Baidu favored by investors, how big a wave do you think DeepMap can create?