Category: Self-driving

Haomo’s self-driving ambitions leap forward with launch of DriveGPT

DriveGPT will reshape the route of vehicle intelligence technology, allowing assisted driving to evolve faster and autonomous driving to come sooner, said Haomo CEO.

(Image credit: Haomo)

The popularity of ChatGPT has led players in the autonomous driving industry to start expecting more from artificial intelligence (AI) technology. Now, a Chinese startup has released a large model of self-driving cognition, becoming the first in the industry to do so.

Great Wall Motor-backed startup Haomo.AI Technology officially launched DriveGPT today, after it previewed the AI large model in February.

DriveGPT is the first generative large model for autonomous driving that will reshape the technological path to automotive intelligence, the Beijing-based company announced today at its 8th Haomo AI Day event.

The AI model builds RLHF (Reinforcement Learning from Human Feedback) technology by bringing in driving data for continuous optimization of self-driving cognitive decision models.

The goal of DriveGPT is to achieve end-to-end autonomous driving, and at this stage, it is mainly used to solve the cognitive decision problem of autonomous driving, said Gu Weihao, CEO of Haomo, at today's event.

DriveGPT has been opened to a limited number of partners, and the first batch includes Beijing Jiaotong University School of Computer and Information Technology, Qualcomm, Volcano Engine, Huawei Cloud, JD Technology, NavInfo, WEY Brand, and Intel.

What is DriveGPT?

GPT (Generative Pre-trained Transformer), essentially solves for the probability of the next word occurring, sampling from a probability distribution and generating a word with each call.

This continuous loop generates a sequence of characters for various downstream tasks. ChatGPT evolved on such a model.

On February 17, as the AI frenzy generated by ChatGPT continues worldwide, Haomo upgraded its large model of self-driving cognition to DriveGPT, calling it the first of its kind in the world of self-driving cognition.

DriveGPT completed model building and validation of the first phase of data with parameters comparable in size to GPT-2, the company said at the time, adding that DriveGPT will continue to introduce large-scale real data next to continuously improve its measurement.

At today's Haomo AI Day, Gu provided more technical details about DriveGPT.

Simply put, DriveGPT is the GPT large model in autonomous driving, which is also solving for the probability of the next Token, according to Gu.

Each call is equivalent to generating a Token based on a sequence of pre-trained Tokens, which are equivalent to a word in natural language processing, except that DriveGPT's Tokens are used to describe driving scenarios.

The sequence of Tokens is a complete time sequence of driving scenarios, including the next possible state of the entire traffic environment and the state of the car.

Under a unified generative framework, DriveGPT can do multiple tasks including planning, decision making and reasoning all together, according to Gu.

DriveGPT's pre-trained model uses 120 billion parameters and is based on 40 million production vehicles' driving data.

In addition to that, Haomo used partially screened human takeover data -- roughly 50,000 Clips -- for training the feedback model, according to Gu.

To easily understand the logic behind this, Gu gives an example:

Suppose we give a prompt to reach a certain target point, this prompt may come from the navigation of a map, or in the future it may come from human voice instructions during voice interaction.

DriveGPT will generate many possible driving styles, some aggressive -- change lanes continuously to overtake and reach the target point quickly, and some will be more steady -- follow other cars to the end.

If there are no other additional prompts, then DriveGPT will follow the optimal results from the feedback training and deliver a result that is more in line with most people's driving preferences.

Coming soon

Haomo is continuing to optimize DriveGPT, and the initial results will first be used in the new Mocha DHT-PHEV with Haomo HPilot 3.0, Gu said.

The Mocha DHT-PHEV is a model of Great Wall Motor's premium brand Wey, and Gu said the model will be available soon.

"We can also provide an excellent driving experience for city NOH (Navigation on HPilot)," Gu said.

In addition to city NOH, scenarios including city cruising, driving shortcut recommendations, and extrication, autonomous driving capabilities are also expected to see new breakthroughs, according to Gu.

"Ultimately we hope to arrive at the ultimate scenario of autonomous driving: driverless," he said, adding, "We believe that DriveGPT will reshape the technological path to vehicle intelligence, allowing assisted driving to evolve faster and autonomous driving to arrive sooner."

Notably, Haomo is one of the strongest players in the autonomous driving space, despite being an under-the-radar startup, especially compared to high-profile local electric vehicle companies.

Haomo was the former autonomous driving division of Great Wall Motor and became an independent company in November 2019. Its core team members come partly from Great Wall Motor's smart driving R&D team and partly from local tech companies.

In September 2022, Funky Cat and Wey Coffee 01 from Great Wall Motor achieved results comparable to the Tesla Model Y in a European safety test.

Both the Ora Funky Cat and the Wey Coffee 01 were equipped with assisted driving technology provided by Haomo, which was a major reason for their high scores, especially on Safety Assist, at the time.

At the 7th AI Day event held today, January 5, Haomo announced MANA OASIS, an autonomous driving computing center launched by it and Volcano Engine, a cloud service platform owned by ByteDance, with a total computing power of 670 PFLOPS, the largest in China at the time.

The computing power support provided by MANA OASIS was key to the training and technology becoming available for DriveGPT, Gu said today.

Separately, Haomo chairman Zhang Kai said at today's event that in addition to the new Mocha DHT-PHEV, the second HPilot 3.0-equipped model, Wey Lanshan, will be launched this year.

Haomo's HPilot has been installed in nearly 20 models and has assisted users to drive more than 40 million kilometers, Zhang said.

In overseas markets, vehicles equipped with HPilot have been shipped to regions including the European Union and Israel and delivered to local consumers.

Vehicles equipped with the driving assistance system will be available in the Middle East, South Africa and Australia, and the Mexican and Russian versions of HPilot will be in mass production soon, Zhang said.

The post Haomo's self-driving ambitions leap forward with launch of DriveGPT appeared first on CnEVPost.

For more articles, please visit CnEVPost.

Chinese self-driving startup WeRide plans to start autonomous vehicle testing in Singapore this year

In September 2022, WeRide launched its Singapore company and signed a memorandum of cooperation with a subsidiary of a local public transport operator in November.

(Image credit: WeRide)

Chinese startup WeRide plans to start testing autonomous vehicles in Singapore this year, as it begins to set its sights on Southeast Asian markets.

WeRide will work with Strides, a subsidiary of Singapore's largest public transport operator SMRT, to test and pilot self-driving vehicles in the Southeast Asian country, according to an article yesterday on its official WeChat account.

In September 2022, WeRide established the Singapore branch and signed a memorandum of cooperation with Strides in November, according to the article.

SMRT and another Singaporean investment house, K3 Ventures, are both strategic investors in WeRide, and they will help the company grow in Singapore and Southeast Asian markets, the article said.

On March 27, Singapore Prime Minister Lee Hsien Loong kicked off a six-day visit to China and visited WeRide's headquarters in Guangzhou, said the article.

WeRide's office in Singapore is hiring talents there, and the company will use the office as a regional development center for Asia Pacific operations, according to its founder and CEO Han Xu.

Founded in Guangzhou in 2017, WeRide's products include Robotaxi, Mini Robobus, Robovan, Robo Street Sweeper and SAE L2-L4 advanced smart driving solutions.

WeRide is now testing and operating driverless vehicles in 25 cities in 5 countries, according to the article.

The company has a fleet of more than 500 self-driving vehicles and more than 16 million miles of autonomous driving, WeRide said.

On March 13, Bloomberg reported that WeRide has confidentially filed for an initial public offering in the US and hopes to raise as much as $500 million.

WeRide is poised to go public as early as the first half of this year, according to the report.

In March 2022, local media outlet LatePost reported that WeRide had closed a new round of over $400 million in financing, with a post-investment valuation of $4.4 billion.

Investors in the round included GAC Group, Bosch, China-Arab Investment Funds, and Carlyle Group, the report said.

WeRide unveils smaller, lower-cost self-driving sensor suite

The post Chinese self-driving startup WeRide plans to start autonomous vehicle testing in Singapore this year appeared first on CnEVPost.

For more articles, please visit CnEVPost.

Alibaba-backed DeepRoute unveils self-driving solution that doesn’t require HD maps

Passenger cars integrated with Driver 3.0 will be available at an established car brand this year, DeepRoute said.

(Car integrated with Driver 3.0 HD map-free driving smoothly in Shanghai. Image credit: DeepRoute)

Many automakers in China are trying to move away from reliance on high-definition maps as they develop technology. Now, a technology provider has a solution that promises to speed up that process.

DeepRoute, a local self-driving startup backed by Alibaba, today unveiled its new Driver 3.0 solution, which it says is a solution that eliminates the need for high-definition maps and can facilitate mass production for automakers.

DeepRoute.ai is one of the first companies to successfully complete public road tests of autonomous driving without HD maps, thus breaking down the limitations imposed by geo-fencing, it said.

The company is also among the first to win production contracts from automakers to produce self-driving cars for consumer use, it said, adding that vehicles integrated with the Driver 3.0 self-driving solution will hit the market in 2023.

DeepRoute shared a video showing lane-level information around the car being generated in real-time without HD maps.

Driving on its own in Shanghai during rush hour traffic, the car interacts safely with pedestrians, e-bikes and others using the road and remains consistent in all road conditions.

It is capable of adaptive cruise control (ACC), stop and go, obstacle avoidance, unprotected left turns, and other technically complex maneuvers.

Driver 3.0 includes two versions of its autonomous driving solution for automakers, D-PRO and D-AIR.

D-PRO costs $2,000 in hardware and includes operations and features that do not require HD maps, such as Valet Parking Assist (VPA), and point-to-point navigation on all roads without operational design domain (ODD) restrictions.

D-AIR costs $1,000 in hardware and focuses on driver assistance that does not require HD maps, such as Automatic Cruise Control (ACC), Lane Centering Control (LCC), and Autonomous Emergency Braking (AEB).

Both versions of the Driver 3.0 solution accelerate automakers’ mass production plans because they comprehensively address concerns about the high cost of mapping and maintenance and the limitations of geo-fencing, according to DeepRoute.

“We focus on rapidly bringing highly advanced, affordable autonomous driving to automobile OEMs,” said Maxwell Zhou, CEO of DeepRoute.

Unlike most other autonomous driving solution providers, DeepRoute is focused on developing an autonomous driving framework and commercializing it by deploying ADAS capabilities first, Zhou said.

DeepRoute strategized the HD map-free approach back in 2020 and began working with OEMs last year based on this approach, he said, adding that without relying on HD maps, smart driving will be available everywhere and affordable for both automakers and consumers.

Passenger cars integrated with Driver 3.0 will be available at an established car brand this year, DeepRoute said, without mentioning the brand’s name.

DeepRoute was founded in February 2019 and became the first company to be able to conduct robotaxi passenger tests in Shenzhen in April 2021.

On September 14, 2021, DeepRoute announced the closing of a $300 million Series B round led by Alibaba, making it the first such company the e-commerce giant has invested in China.

The high cost of self-driving kits is one of the factors preventing the technology from being used at scale, and one of the goals of DeepRoute’s efforts over the past few years has been to bring the cost down.

On December 8, 2021, DeepRoute unveiled Driver 2.0 for under $10,000, nearly the same price as FSD and the lowest recorded in the industry. Driver 2.0 was available in brands including Lincoln and ‘s Geometry at the time of launch.

On April 20, 2022, DeepRoute announced its first fleet of 30 vehicles with the Driver 2.0 solution, which will be put into the company’s robotaxi operation in Shenzhen.

Chinese self-driving startup DeepRoute unveils L4 self-driving solution that costs less than $10,000

The post Alibaba-backed DeepRoute unveils self-driving solution that doesn't require HD maps appeared first on CnEVPost.

For more articles, please visit CnEVPost.

Baidu’s robotaxi platform Apollo Go gets permit to offer fully driverless rides in Beijing

To date, Baidu's Apollo Go has been offering fully driverless ride-hailing services in Beijing, Wuhan and Chongqing.  | Baidu.US | Baidu.HK

(Image credit: Baidu)

Baidu's ride-hailing service platform Apollo Go has won a permit to offer fully driverless rides in Beijing, allowing it to expand the service to three Chinese megacities.

The search engine giant announced the development today, saying it is the first provider of fully driverless robotaxi services in the capital city of any country worldwide. Apollo Go has previously been approved to offer the service in Wuhan and Chongqing.

Baidu Apollo will deploy a total of 10 fully driverless vehicles in Beijing's Yizhuang Economic Development Zone, according to a press release from the company.

Apollo Go is currently providing an average of more than 20 rides per vehicle per day within the area, exceeding the average number of rides taken by traditional online ride-hailing services, Baidu said.

Yizhuang is one of the active hubs for autonomous driving in China. Beijing plans to expand its high-level automated demonstration area in the Yizhuang Economic Development Zone from the existing 60 square kilometers to an eventual 500 square kilometers.

Baidu has been developing autonomous driving technology since 2013 and has accumulated more than 50 million kilometers of testing in Level 4 autonomous driving.

As of the end of January, Apollo Go offered more than 2 million cumulative rides to the public, Baidu said.

In the fourth quarter of 2022, Apollo Go provided 561,000 rides to the public, up 162 percent year-on-year, according to Baidu's fourth-quarter earnings report.

On November 29 last year, Baidu announced that it plans to scale up Apollo's operations in 2023 with fully unmanned self-driving operations in more regions.

Baidu will build the world's largest fully driverless taxi service area in 2023, maintaining its growth momentum as the world's largest robot cab provider, the company said at the time.

Baidu previously announced plans to expand its self-driving mobility service to 65 cities by 2025 and 100 cities by 2030.

Baidu plans to put 200 additional driverless vehicles into operation in 2023

The post Baidu's robotaxi platform Apollo Go gets permit to offer fully driverless rides in Beijing appeared first on CnEVPost.

For more articles, please visit CnEVPost.

driive.com
All about Smart and Sustainable Mobility

Archives

December 2024
M T W T F S S
 1
2345678
9101112131415
16171819202122
23242526272829
3031