"Autonomous driving" is something for the 2030s and later, is the common opinion among the German public after the hype died down a few years ago.
Not so in China! First of all, the state has a long-term plan 'Made in China 2025' with a strong political will to take over the global leadership of the automotive market, or rather the mobility market. The fact that this will lead to considerable losses in jobs and importance for the previously leading German automotive industry and its suppliers has been clear since the politically enforced shift to e-mobility in the EU at the latest.
However, the Chinese car buyer also plays a key role here. On the one hand, there is the newfound pride in their own superior New Electric Vehicles (NEV). Added to this is the need to spend time on longer journeys to work in a more useful way than driving.
Autonomous cabs in China collect data
The Chinese internet platform Baidu has been operating autonomous cabs since 2020, currently in ten cities from Wuhan and Shenzhen to Beijing. There are now more than a thousand 'Apollo Go' robot cabs in use in China, more than 300 in the city of Wuhan alone with its 11 million inhabitants (Fig. 2). This year, 24-hour operation was also approved in Wuhan. Since August 2023, there has been a permit for driving on the highway to Wuhan Airport. Since February of this year, the autonomous airport service has also been available in Beijing to Daxing Airport. The data from over 6 million journeys by private cars by the beginning of 2024 is a valuable database for advancing autonomous driving towards Class 4.
Fig. 2: Driverless robot cab 'Baidu Apollo Go' in Shenzhen
Tesla drives the market in China
Now that Tesla has lost the market leadership in NEVs in China to BYD and price reductions have also faded, automated driving is the next challenge. For a long time, Tesla was the self-proclaimed technology leader in semi-autonomous driving. The 'Full Driving System' FSD, initially only based on cameras and dispensing with expensive LiDAR sensors, has now been offered to Chinese car buyers as a beta version. In international nomenclature, this corresponds to level 2+, in which the driver can let go of the steering wheel but must take over the wheel again at any time (usually within ten seconds).
Xpeng, Li Auto and Huawei are the NOA frontrunners
NOA (Navigation on Autopilot) is the Chinese terminology for (semi-)autonomous driving. Li Auto plans to offer its semi-autonomous driving system for 100 Chinese cities by the end of 2024. Xpeng plans to deploy its AI-supported car operating system in 200 cities. For Xpeng, AI Smart EVs are characterized by three core features: active learning, fast growth and personalized experience. The core technology is to feed and train large AI models with unlimited situations to improve their understanding, perception and decision-making ability in complex scenarios. Currently, Xpeng has developed core algorithms for perception, positioning, planning and decision making. This is complemented by both in-vehicle and cloud-based data processing capabilities that enable rapid iteration of algorithms based on real-world data and provide users with continuously improved autonomous driving capabilities via OTA updates, as Xpeng recently announced at a press conference. With the AI models, Xpeng aims to achieve highly automated driving class L-4 by 2025.
In June of this year, the P5 model with X-Pilot version 3.5 was launched on the market. Five radar sensors, two LiDAR systems, 12 ultrasonic systems. Combined with the AI-based software, the P5 is not only capable of highly automated driving on the highway, but also in urban areas.
Fig. 3: The P5 model from Xpeng (Xiaopeng Motors, Guangzhou) was launched on the market in June 2024 with X-Pilot version 3.5 for Level 3 automated driving
Automated driving Mercedes
In the first half of 2024, Mercedes sales in China shrank by 9% to 342,000 vehicles. The operating result across all markets plummeted by 25% to €7.9 billion compared to the previous year. No wonder that Mercedes is trying to get involved in fully automated driving level 4 in China, as the test vehicle in Figure 1 shows. With the familiar combination of LiDAR, camera, radar and ultrasonic sensors, linked to a high-performance computer, Mercedes has currently achieved Level 3 approval in Nevada, California and Germany. The production-ready version of the DRIVE PILOT will be installed in the S-Class and the EQS. In Germany, highly automated driving is permitted on certain highways up to a speed of 60 km/h, in Nevada and California up to 40 mph (64 km/h). From the end of 2024, the speed is to be increased to 90 km/h and, according to Mercedes, to 130 km/h by the end of this decade. Mercedes proposes that all automated vehicles should be identified to other road users by turquoise marker lights.
The stages to the self-driving car
Levels 1-3 are currently being used on the market, with level 4 in preparation.
Level 0 refers to non-automated driving, in which the driver performs all driving functions themselves, even if support systems (e.g. ABS) are present.
Level 1: In so-called 'assisted driving', only certain assistance systems (e.g. dynamic adaptive cruise control) help to operate the vehicle.
Level 2: In 'semi-automated driving', many functions such as parking, lane keeping, general longitudinal guidance, acceleration and braking are taken over by the system (e.g. the traffic jam assistant), but the driver must keep a constant eye on this.
Level 3: In 'highly automated driving', the hands no longer need to be on the steering wheel. Here too, however, the system independently takes over many driving tasks such as braking, steering, changing lanes or overtaking. The driver can turn their attention to other things, but is prompted by the system if necessary, usually within 10 seconds, to take over the lead.
Level 4: In 'fully automated driving', the system also takes over control of the vehicle (lateral and longitudinal control) permanently. However, the driver can be asked to take over if the driving tasks can no longer be handled by the system. However, the system is able to return the vehicle itself from any initial situation to a system state that minimizes risk (e.g. by bringing the vehicle to a standstill on the hard shoulder) if the driver does not take the lead.
In a nutshell
- China has a master plan to achieve technological dominance and market leadership. Key technologies according to the 'Made in China 2025' strategy include electromobility and AI
- China's electric cars are AI on wheels, while European cars are still trying to become intelligent. The combination of 'software + battery + base chassis' allows tech companies like smartphone maker Xiaomi to enter the business and quickly develop new cars.
- With driverless 'Baidu Apollo Go' robotaxis, valuable data for autonomous driving is being generated from 6 million journeys to date.(Fig. 2)
- With the P5 model from Xpeng in Guangzhou, the X-Pilot version 3.5 for level 3 automated driving was recently launched on the market(Fig. 3). Other manufacturers such as Li Auto and Huawei are also working hard on technologies for Level 3 and 4.
- Mercedes is trying to stay on the ball and is involved in Level 4, fully automated driving, with test drives in Beijing.(Fig. 1)
What do you need for a successful, growing economy? First of all, education, i.e. good schooling and training, especially in STEM subjects, i.e. the natural sciences. This results in the availability of qualified specialists, but also in a society that is friendly to progress. And a government and authorities that see themselves as supporters of ideas for the future and the further development of their country, rather than obstructors. Unfortunately, we lack most of these things. It's no wonder that Germany is falling further and further behind other countries and that jobs and prosperity are being lost as a result.
Please stay with us.
Best regards
Yours
Hans-Joachim Friedrichkeit
