Electronic Enthusiasts Network reported (text / Li Wanwan) Robotaxi, a self-driving taxi, is a vehicle that is based on self-driving technology and can provide rental services. According to research and consulting agency IHS Markit, by 2030, the total market size of China's shared travel will reach 2.25 trillion, of which Robotaxi will account for 60%, with a scale of 1.3 trillion.

At present, self-driving taxis have been operating in many cities. Among them, Baidu Apollo has become the world's largest self-driving travel service provider. The company can also provide intelligent solutions for other OEMs. In addition, Pony.ai, WeRide, Yuanrongqi Autonomous driving technology companies such as Xingxing are also cooperating with car companies and travel service providers in the Robotaxi market.

Baidu Apollo

Baidu began to deploy autonomous driving in 2013. In 2017, it launched the world's first open platform for autonomous driving, Apollo. At present, Baidu Apollo has leading solutions in the three fields of autonomous driving, intelligent driving and intelligent transportation.

In terms of self-driving travel, Baidu’s Robotaxi platform, Radish Kuaipao, has been launched in Beijing, Shanghai, Guangzhou, Shenzhen, Chongqing, Wuhan and other cities. In the first quarter of this year, Radish Kuaubo provided 196,000 rides.

Baidu Apollo can provide automotive intelligent solutions for other auto OEMs. In response to problems such as complex parking scenarios, shortage of resources in domestic popular business districts and no complete solution, Baidu Apollo launched the L4-level automatic parking technology AVP system and adapted it to passenger cars.

AVP relies on 1 monocular wide-angle (120°) camera, 4 ultra-wide-angle (190°) cameras, 8 UPA ultrasonic sensors, 4 APA ultrasonic sensors and other equipment, with Baidu's high-precision map and decision-making core ACU's optic nerve The system, whether it is a horizontal or vertical parking space, or a space parking space that is inclined and has no markings, can achieve skilled automatic parking.

In addition, Baidu Apollo's ANP parking/high-speed/city three-domain integration product, based on Apollo Lite vision technology, can connect to parking through L4 technology's dimensionality reduction, high-precision maps, and light sensor and light computing power production solutions domain and vehicle domain, providing users with assisted driving capabilities with close to zero intervention from the starting point to the end point.


Pony.ai was established at the end of 2016. In January this year, Pony.ai released its sixth-generation autonomous driving system. This system is equipped with 23 sensors, 4 solid-state lidars and 7 cameras on the roof, and 5 cameras around the body. 1 millimeter-wave radar and 4 cameras, and it was also equipped with Orin, the industry's strongest autonomous driving SoC at the time.

On April 24 this year, Pony.ai announced that it has won the bid for the 2022 taxi capacity index in Nansha District, Guangzhou. On April 26, the company announced further cooperation with Ruqi Travel, a mobile travel platform under GAC Group, to integrate the two parties in technology research and development and travel. Resources in the service field, build a fleet of 100 autonomous driving vehicles, and plan to start the Robotaxi demonstration operation in Guangzhou within this year.

On June 22, Pony.ai announced that it has officially delivered samples of the Autonomous Driving Controller ADC (Autonomous Driving Controller) based on the NVIDIA DRIVE Hyperion computing architecture and DRIVE Orin system-on-chip. The fourth quarter of 2022 will officially open.

The Pony.ai autonomous driving domain controller has two versions, single-orin and dual-orin, with a computing power of 254TOPS and 508TOPS respectively. As the world's leading automotive-grade SoC, NVIDIA DRIVE Orin can handle a large number of applications and deep neural networks simultaneously, while meeting system-level safety standards to ensure safe, advanced, and intelligent vehicle performance.

WeRide know-how

WeRide was established in 2017 and has L4-level autonomous driving solutions. In 2019, it launched China's first fully open Robotaxi service to the public, and began normal operations in Guangzhou. WeRide has reached strategic partnerships with many of the world's top OEMs and Tier 1 suppliers, including Renault-Nissan-Mitsubishi Alliance, Yutong Group, GAC Group, Bosch, etc.

Not long ago, the company released a new generation of autonomous driving sensor suite WeRide Sensor Suite 5.0 (WeRide SS 5.0 for short), and announced that the Robotaxi fleet equipped with the suite has been fully put into large-scale testing and applications.

WeRide SS 5.0 has 12 cameras and 7 solid-state lidars, which together form 6 perception modules, which are cleverly arranged in different positions of the body, so that the automatic driving kit can truly integrate with the car, and maximize the preservation of the original shape of the body. All-round powerful perception ability.

Thanks to the further optimization of the solid-state lidar and computing power platform, WeRide SS 5.0 can run stably on the computing power platform of 250-500TOPS, further reducing the computing power threshold of Robotaxi.

Yuan Rong Qixing

Yuanrong Qixing focuses on the research and development and application of L4-level autonomous driving technology, focusing on two types of business, travel and intra-city freight, and has two product lines, Yuanqixing (Robotaxi self-driving passenger car) and Yuanqiyun (Robotruck self-driving light truck).

Yuanrong Qixing has R&D centers in Shenzhen and Beijing, and has carried out autonomous driving-related cooperation with well-known enterprises such as Dongfeng Motor and Cao Cao Travel in many places across the country, and has conducted a large number of road tests and trial operations. On April 20, 2022, Yuanrong Qixing's L4-level autonomous driving pre-installation solution fleet landed in Shenzhen. The fleet consists of 30 autonomous vehicles.

The fleet is equipped with DeepRoute-Driver 2.0, the company's latest L4-level autonomous driving solution for front-loading mass production. In terms of performance, DeepRoute-Driver 2.0 can calmly complete complex operations such as changing lanes and avoiding temporary roadblocks, solving special road conditions such as ghost probes and jamming, and can be adapted to a variety of models. , Geely Geometry A and other models have been successfully adapted, and you can enter the automatic driving system with one key.

Yuanrong Qixing's pre-installation solution, each car's body is embedded with 2 to 5 solid-state lidars and 8 cameras, and the sensors have passed the vehicle-grade high temperature and waterproof detection. In the future, its sensor configuration can also be adapted to different sensor configurations according to the design requirements of car manufacturers. In terms of computing power, DeepRoute-Driver 2.0 is equipped with NVIDIA DRIVE Orin automotive-grade chips, combined with Yuanrong Qixing's self-developed inference engine technology, to control the overall power consumption of the computing platform to 150 watts.

Prospects and challenges for Robotaxi

Drivers of traditional taxis and ride-sharing services will try to avoid short-distance passengers as the fare is small, and robo-taxi does not have this distinction, so robo-taxi is expected to bring huge business opportunities for taxi and ride-sharing service providers . At present, labor costs account for about 60% of the total cost of Chinese taxi companies, and the proportion of driver costs in carpooling costs is as high as 80%.

However, the challenge is that self-driving cars are expensive. For example, Baidu's self-driving cars cost 480,000 yuan each. Pony.ai's Robotaxi is based on Lexus' RX 450h. Driving system, the cost may reach about 1 million.

Moreover, the current Robotaxi also needs to be equipped with a safety officer. The overall cost is equivalent to adding the driver's salary. Therefore, at present, the cost of the Robotaxi is very high, but the follow-up car can be changed to a cheaper model, and then with the automatic driving As the technology is more mature, Robotaxi can gradually eliminate the need for security officers, but this may take a long time.

Leave a Reply

Your email address will not be published.