This afternoon, a small conference room on the first floor of the Haval Research and Development Centre was the site of a major announcement by Great Wall Automobile. The company unveiled its latest achievements in the field of intelligent technology and outlined the future direction of their brands in terms of intelligence development.
Especially noteworthy is the company’s plan for intelligence after the establishment of the AI Lab. How to utilise AI to achieve superior intelligent cockpits, assisted driving and service ecology was the key focus of the announcement.
Application of Large-Scale Models
As a brand representing Great Wall Automobile’s integrated intelligent technology, the suite called “Coffee intelligence” encompasses four major intelligent products: Coffee Intelligent Cockpit System, Coffee Intelligent Driving System, Coffee Intelligent Cloud, and Coffee Electronic Architecture.
Large-scale AI models are mainly used in intelligent cockpits and assisted driving, which Great Wall describes as Space GPT and Drive GPT.
Regarding intelligent cockpits, Coffee OS 2 has already been incorporated into the second-generation Haval Big Dog, DS Dragon MAX, and Velociraptor, and WEY brand’s new Mocha, Blue Mountain, and High Mountain models.
In our practical experience, these vehicles have a highly unified intelligent cockpit system. Whether it is UI design or App layout, they can adapt to screens of different sizes and aspect ratios.
But that’s not nearly enough.
In terms of ecology, Great Wall plans to fully open the stack, beyond the existing application store. By integrating mobile application stores and enabling all mobile applications to be opened in the car, they’ll offer a combined car-phone mini-programs and integrate consumer electronics, AIoT devices, and docking hardware. This outlines the next phase of Great Wall’s planning in the intelligent cockpit field.
Another key area of intelligence focus within the cockpit is voice recognition. At present, the biggest pain point with voice assistants is that they don’t mimic human conversation well. If I suddenly forget what the “sunshade” is called, can I just point and have it opened for me?
Multimodal perception is the key. By responding to voice, vision, and gestures, interaction within the cockpit can be more natural, efficient, and intelligent.
Coupled with proactive services, vehicle-wide SOA architecture, and the use of Space GPT, Great Wall’s intelligent cockpits will be dramatically smarter.
From Intelligent Cockpit to Intelligent Space
The other core area for intelligence transformation is assisted driving. At present, Great Wall’s Coffee Pilot has been implemented into three product generations and is integrated into nearly 20 models.
The upcoming period of competition will likely focus on navigational assisted driving, specifically urban navigational assisted driving.According to the schedule, urban NOH of Great Wall Automobile is set to expand across a hundred cities next year.
Simultaneously, high-speed/expressways without graphical schemes, urban fixed scenes, and the urban NOH will also be updated in succession.
The process of assisted driving has its challenges. Great Wall Automobile shared an example where individuals with small eyes are often misidentified by Driver Monitoring System (DMS) as dozing off. How could such a “simple” matter pose a problem?
In most solutions, DMS cameras only examine the opening and closing of the eyelids to judge whether the driver is cognitively distracted, leading to significant errors, particularly among users with relatively small eyes where misjudgments occur more frequently.
Great Wall’s solution is straightforward: if one method doesn’t work, run multiple checks using a multi-modal algorithm framework. The DMS camera then resolves the issue by evaluating the driver’s cognitive status based on a combination of eyelid openings, blink rate, yawning, and other actions.
Today, intelligence is the most competitive area in the automobile industry. Screens, chips, software, none can be overlooked.
Great Wall Automobile, as one of the core forces among independent brands, has demonstrated its strength in the field of smart automation.
This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email email@example.com.