Jiyue 01 pure visual high-end intelligent driving firmly stands in the first echelon V1.4.0 upgrades OCC 3D map visualization effect

In different eras, cars have different forms. Nowadays, cars are no longer a simple means of transportation. AI technology has become the "soul" of cars, making cars like "Transformers" with dual identities of travel stewards and close friends in life. It can not only act like a smart and capable robot butler, autonomously completing various driving tasks, it can also transform into the car owner's emotional assistant, playing games, watching TV series, chatting freely, and changing its tricks to bring the car owner a pleasant journey experience both physically and mentally.

Jiyue Automobile Robot is based on the ecological empowerment of Baidu AI large model and has ushered in a new evolution. On March 25, as the industry's first automotive AI technology conference, Jiyue held AI DAY 2024 in Beijing. Jiyue CEO Xia Yiping released a new version of OTA V1.4.0 software. Baidu's three most cutting-edge AI native technologies will continue to support it. Extreme car robot evolution. Thanks to world-class leading AI technology empowerment, Jiyue will set an industry benchmark in the AI ​​travel era.

The V1.4.0 version is also the third consecutive OTA launch of Jiyue 01 within 90 days and the fifth OTA since its launch. This upgrade involves many fields such as smart driving, smart cockpit, smart interconnection, and electricity and gas. A total of 200 functions have been upgraded and 5,300 user feedback issues have been resolved.

For example, the OCC occupied network visualization effect has been upgraded, the two-way integration of SIMO and Wen Xinyiyan's large model, and the human, car, and home ecology based on the large model have been officially "launched", etc. The V1.4.0 version will bring users a smart travel experience that is better to drive, better to use, and more fun, and once again fulfills the promise of allowing users to "drive a new car every month."

Baidu AI large model ecology is deeply empowered, Jiyue automotive robots accelerate the evolution

The automobile revolution has experienced the "fuel vehicle 1.0 era" and "electric vehicle" 2.0 era", and in the current "smart car 3.0 era", the underlying driving force comes from AI large model technology. The deep integration of AI large models and automobile manufacturing will lead to a new generation of automobile product revolution.

The era of smart cars 3.0 is the era of automobile robots. The starting point of the change of the era is the transfer of the "driving rights" of automobiles from humans to AI. Jiyue Automotive Robot has clearly defined the product characteristics of the 3.0 era, which should meet the core capabilities of the "Three Self".

(Jiyue CEO Xia Yiping)

Since the birth of the Jiyue Automobile Robot, it has taken "natural communication, free movement, and self-growth" as its core product concepts, aiming to enable smart cars to think and act like humans and continue to evolve. Xia Yiping said: "In the era of smart cars 3.0, 'robotization' will be the development trend. Cars with the three abilities of natural communication, free movement, and self-growth are car robots and are the true 'AI CAR'." Jiyue Auto Robot allows everyone to see what smart cars really look like, and feel the technological innovation, efficiency improvement, and experience subversion brought by AI."

As a high-end smart car robot brand, Jiyue Auto Robot has Geely SEA's vast architecture and ecological empowerment of Baidu's AI capabilities provide top-notch mechanical literacy and the industry's first-class intelligence level. Thanks to the ecological support of Baidu AI large model, Jiyue 01 has deeply integrated the technical capabilities of Wen Xinyiyan, Apollo, Baidu Maps, Xiaodu and other technologies. Jiyue 01 has created many industry leaders such as the first in the world to realize the "on-car" AI large model, the only pure visual high-end intelligent driving in China/the only one in the world, the first OCC occupied network upgrade, the first full localization of AI voice algorithm, etc.

In this AI DAY, Baidu’s three latest AI native technologies in the fields of autonomous driving, LD maps, and human-computer interaction were unveiled and will be applied to Jiyue Automobile Robots.

In terms of high-end intelligent driving, based on Baidu Apollo's pure visual high-end intelligent driving capabilities and safety system empowerment, Jiyue has completed the OCC occupied network upgrade and has formed a complete "B.O.T" technology system. At this AI DAY, Baidu Apollo autonomous driving vision model VTA (Vision Takes All) was officially released. It has been significantly upgraded to include dynamic and static detection, timing tracking, real-time mapping, scene understanding and other capabilities, once again improving the extremely pure visual high-end intelligent driving. Capability limit.

(Wang Liang, Chief R&D Architect of Baidu Intelligent Driving Business Group and Chairman of IDG Technical Committee)

Baidu Intelligent Driving Business Wang Liang, chief R&D architect of the group and chairman of the IDG technical committee, said: "Based on large models, Baidu has created the industry's first smart driving data production line and LLM-empowered autonomous driving data index. At the same time, through generative AI technology , Baidu also has the ability to efficiently process long-tail data, which are important data engines that promote the development of end-to-end autonomous driving technology."

Regarding intelligent driving maps, "with pictures" and "without pictures" have become a hot topic in current intelligent driving. Baidu Maps LD map (LD stands for Lane Level Navigation) is a native map for autonomous driving, which is generated from a large visual perception model for autonomous driving. It completely reconstructs the traditional map production model, realizes a new paradigm of end-to-end map generation, and breaks through the lane-level map. The bottleneck of scale. LD was born for autonomous driving. It minimizes the elements and accuracy of traditional high-precision maps and retains the most necessary elements and accuracy. At the same time, it adds experience layers, safety layers, real-time layers and other traditional high-precision maps. Elements that maps do not have make autonomous driving safer.

(Baidu Vice President Shang Guobin)

Jiyue 01 fully integrates the core capabilities of Baidu Maps. As long as Baidu Map Navigation can Wherever you go, Jiyue 01 PPA smart driving can be used. Shang Guobin, vice president of Baidu, said: "LD Map has three characteristics: 'can be used in cities across the country', 'can be used on mobile phones and cabin driving', and 'intelligent driving only requires maps to be safe'. It truly covers 360 cities across the country. The world's only company that applies autonomous driving native maps to mobile phones and cars, allowing smart driving to drive like a human and safer than humans."

At the voice interaction level, in order to allow the car robot to interact with the user Smarter and faster. Based on the underlying capabilities of Baidu's voice interaction, Jiyue 01 has built the voice assistant SIMO into a voice ceiling with the "all-area and all-end-side" in-vehicle voice. The recognition speed is as fast as 500ms, and it can interact smoothly in weak/no network environments.

In the future, multi-modal interaction solutions for video and voice fusion will be launched soon, comprehensively improving multi-modal fusion interaction capabilities and user privacy protection, and further improving speech recognition accuracy through an all-in-one end-side model.

(Jia Lei, Chief Architect of Baidu Voice)

Jia Lei, Chief Architect of Baidu Voice, said: “More powerful The technology must be a terminal-side large-scale model, which allows the vehicle to quickly respond to user interaction needs and answer vehicle-related operation questions in any network environment, bringing users a perfect terminal-side voice interaction experience. Video and voice fusion, multiple The integrated and terminal-side large models are industry-leading revolutionary technologies, which will all be used to support the interactive upgrade of Jiyue automotive robots."

PPA added Guangzhou Kaicheng , OCC 3D map enables perception recognition visualization

Software innovation will bring more surprises to smart cars. Car robots, which have always focused on rapid upgrades and iterations, will bring interesting "souls" to consumers.

As an important part of Jiyue's pure visual intelligent driving solution, the OCC occupied network has been upgraded again, and the perception ability has been greatly improved. On the basis of centimeter-level 3D model descriptions comparable to lidar, the types of obstacle recognition have increased again, including not only single static obstacles such as construction signs, fences, roadblocks, and anti-collision barrels, but also temporary road construction fences, Obstacles such as broken down vehicles parked on the side and temporarily stacked large trash cans can also be accurately sensed, thus once again improving the adaptability of Jiyue Pure Vision high-end intelligent driving during driving.

It is worth mentioning that in order to allow users to better understand the principles of OCC technology and use it with more peace of mind, the V1.4.0 version has added OCC 3D map easter eggs. Users can clearly see the world in the eyes of the car robot through the car screen.

Like building blocks, squares of different colors mark obstacles/green belts/other vehicles outside the car, and cartoon 3D characters mark passing pedestrians, adding a lot of fun to the obscure AI technology. This function also supports turning on and off through voice, which is very convenient to control.

Thanks to the improvement of OCC’s perception capabilities, the point-to-point pilot assistance PPA has also been greatly upgraded, which can handle more complex driving scenarios. For example, at an intersection with complex road conditions, if the car in front turns too slowly or the battery car is too close, PPA can actively detour to improve efficiency.

In the event of sudden road construction fences, roadside broken down vehicles, temporary piles of debris, etc., PPA can flexibly bypass or stop in time. In view of the problem that many users miss the exit of the highway ramp due to long solid lines or large vehicles blocking the road, PPA can also reasonably plan the route and find the right time to change lanes, so that every extreme user can avoid taking the wrong path.

In addition, Urban PPA has opened a new city - Guangzhou, which is also the fifth city after Beijing, Shanghai, Hangzhou and Shenzhen. For the more special urban road sections in Beijing, PPA has also been further improved and expanded its coverage to fully cover the main roads in the inner city of the Fifth Ring Road.

Currently, nearly 90% of users are using Jiyue PPA smart driving, and the cumulative mileage has exceeded 1.8 million kilometers, which is equivalent to circling the earth 45 times. Next, the mileage covered by PPA smart driving will exceed 500,000 kilometers. In 2024, based on the AI ​​large model capabilities of Baidu Apollo and Baidu Maps, the goal of PPA smart driving to be "available nationwide and have a good experience on every road" is just around the corner. As long as Baidu Maps can navigate to places, PPA smart driving can be used.

Jiyue 01 is the world's first smart car to support voice-controlled car/parking outside the car. It has industry-leading valet parking AVP and automatic parking assist APA functions. The APA penetration rate reaches 98%, making it the most advanced car on the planet. Forced parking. In the V1.4.0 version, the parking experience has been greatly upgraded, with new functions such as searching for parking spaces on open roads, active avoidance of passing vehicles during parking, and automatic locking after parking. The level of care and safety awareness are comparable. Five-star parking attendant.

The V1.4.0 version has also upgraded a series of driving experience optimizations, such as: new automatic turn signal function, new PPA high-speed elevated lane change style settings, automatic steering image assistance, new blind spot warning, etc., and optimized Experience ADAS functions such as forward collision warning FCW and automatic emergency braking AEB. In the future, urban PPA driving style selection will also be updated.

Interaction is more natural, entertaining and immersive. The human-car home ecosystem based on large models is officially launched

Nowadays, more and more people want cars to be like "mobile living rooms". The rich and colorful functions meet the needs of almost every family member.

Thanks to the underlying capabilities of Baidu’s AI large model, SIMO has achieved two-way integration with Baidu Wenxinyiyan, bringing a more natural, smooth and logical human-computer interaction experience. In the V1.4.0 version, SIMO has transformed into an all-round butler, adding active recommendations based on large models and free questions and answers about car usage, and supporting functions such as music recommendation, video information Q&A, attraction recommendation and inquiry, and functional answers. If you have any questions, just ask SIMO. The level of natural communication is taken to a higher level, forming a clear generational difference from the rigid and mechanical voice assistants on the market.

For example, if you are chatting with your friends about a TV plot in the car but cannot remember the details, you can quickly find the plot by asking SIMO and play the corresponding episode. . How to use PPA smart driving and other car-related questions can also be easily answered through SIMO.

In addition, a new sharing function for content generated by large models has been added. SIMO supports sending generated text, pictures and other content to Jiyue APP with one click, making it convenient for users to view and share. The SIMO experience outside the car has been further optimized, adding continuous voice dialogue outside the car and improving the recognition accuracy.

#SIMO’s powerful capabilities have become the industry benchmark. Currently, SIMO's daily average voice interaction penetration rate reaches 98%, and the average daily number of voice interactions is as high as 63, which is 10 times that of its industry peers. Among them, each vehicle calls Wenxin Yiyan service 7.1 times per day on average, accounting for approximately 11% of the average number of daily voice interactions.

Rapid iteration and efficient closed-loop enable automotive robots to continuously evolve. In the five months since the launch of Jiyue 01, 23,323 user feedbacks have been received through SIMO. Each problem will be connected to the R&D center with second-level efficiency, and users will be fed back within 24 hours to achieve rapid upgrade and iteration.

In terms of the interconnection of people, cars, and homes that users are generally concerned about, Jiyue 01 has achieved interconnection and interoperability with the Xiaodu ecosystem, which means that the people, cars, and homes ecosystem based on large models has officially been launched. Users can use Xiaodu speakers to complete operations such as vehicle status inquiry, remote vehicle control, and abnormal event reminders at home, and grasp the real-time status of their car at will.

A more user-friendly function upgrade is that the V1.4.0 version also adds an "all-weather extreme mode", which has a wider range of uses and functions than the general camping mode. More powerful. Turning on the all-weather extreme mode can keep the vehicle powered on at all times, and the air conditioning, entertainment services and other functions can be used normally. It is especially suitable for use in scenarios such as camping, sleeping in the car, and temporarily leaving pets in the car, so as to bring users A more comfortable travel life experience.

Multimedia entertainment ecological services have always been a feature frequently used by users. Version V1.4.0 adds a new wireless screen projection application that supports screen projection to the car for playback; a new in-car Wi-Fi hotspot is added for in-car devices to share traffic. At the same time, the games "All-Star Streetball Party" and "Little Triangle Big Hero" were also launched.

In addition, it also includes new app stores, new Baidu network disks, new App remote car control and other functions. Charging experience, steering wheel feel, sentry mode, audio experience, etc. have also been optimized and upgraded. Whether returning home from get off work, waiting to pick someone up, or going camping, Jiyue owners can spend a fulfilling, interesting and comfortable time in the car.

Xia Yiping predicted that the Jiyue V2.0 version will be released soon, and will be upgraded with a new voice model, PPA smart driving that can be opened nationwide, and other major features.

Looking to the future, AI technology will surely create a smart world that subverts imagination. Based on the ecological capabilities of Baidu AI's large model, Jiyue will bring users a more intelligent and comfortable car robot life, and will also help the industry accelerate evolution and lead the era of car robots.

The above is the detailed content of Jiyue 01 pure visual high-end intelligent driving firmly stands in the first echelon V1.4.0 upgrades OCC 3D map visualization effect. For more information, please follow other related articles on the PHP Chinese website!