Great Wall Motors selects L3 LiDAR supplier, opting for pure solid-state technology

When we first contacted Great Wall two years ago, they made such a request. Therefore, the parameters of IbeoNEXT are completely designed to match the autonomous driving function. Its future cost can be predicted. It will be cheaper than mechanical ones. As long as the volume is increased, we think the cost will not be a problem.

Q: OEMs also have corresponding testing links. In comparison, what is the difference in the positioning of Liangdao Smart?

Ju Xueming: In the traditional automotive industry, there are various specialized testing and verification service providers for testing and verification, such as those specializing in engines and electronic modules. Especially in Germany, there are many engineering service companies with thousands or even tens of thousands of employees to undertake many OEM testing and verification tasks.

In fact, we believe that Liangdao Intelligence is targeting a new market. Since there is no testing for intelligent sensing modules in the traditional industry, this is currently a new component in the industry and is a test and verification of intelligent components.

Compared with OEMs and Tier 1, our own positioning is to help OEMs and sensor suppliers, and the three parties work together to improve the safety and reliability of autonomous driving. Because if our tool chain is powerful enough, we can have a more objective feedback on KPIs in a shorter time. This value feedback to OEMs or sensor suppliers is actually an acceleration engine for its algorithm iteration. In this way, the three parties work together to achieve the mass production of autonomous driving. As for the ownership of the data, it needs to be determined based on the specific project.

Q: What specific areas does the Liangdao Intelligent team’s R&D work focus on?

Ju Xueming: Algorithm development based on various sensors, development of tool chains, including the establishment of a big data center.

Q: What are the barriers for testing and verification service providers?

Xueming Ju: The core entrepreneurial team also had relevant experience in German OEM and Tier 1, especially in mass production and development of LiDAR. We have the best understanding of sensors themselves. In addition, we have been targeting this market since 2017. We believe that the mass production of autonomous driving in the future must be tested and verified, so we have been building our tool chain since 2017, whether it is to provide automated annotation or scene capture. Now this tool chain can serve the Great Wall project.

The barriers mentioned here actually require extremely high comprehensive capabilities, even as high as those of algorithm development suppliers. Building this test system is the same as building an environmental perception hardware system and supporting algorithms. You need to understand the normalization and standardization of data collection for the entire test, and you need to work with the OEM and suppliers to complete it.

At the same time, the industry has also put forward new requirements. The current laser radar, especially Flash laser radar, generates a large amount of data per unit time. We need to collect a large amount of mileage data, establish a data platform and data center, and apply various rich tool chains in the data center to perform automatic labeling and automatic scene capture.

Of course, a rich tool chain is also one of the core barriers to meet the requirements of different customers and different scenarios in the future. We hope to establish a complete tool chain and finally develop data mining products based on data.

Q: How is the autonomous driving test verification of multiple vehicles conducted?

Ju Xueming: At present, the test verification of our current products is mainly based on bicycles. We use bicycles as carriers, equipped with sensors to be tested and evaluation systems, to collect data in traffic or other scenarios we need. We are also considering the future model. We have established a team to consider adding various sensors to the facility side to establish the true value. This is equivalent to capturing all the information in the traffic flow as long as you can monitor it. This is more used for driving behavior analysis in the future.