In 2022, the self-driving track has been highly recognized by capital for market segments, and financing information has come one after another; At the same time, the high-level self-driving mass production has a long cycle, high complexity, and the development process and standards are not uniform, which makes the self-driving technology companies face many challenges in promoting mass production. While the auto-driving /ADAS function has gradually become the standard of newly listed models of major car companies, the integration of parking and parking has become a new "window" for car intelligence. Although the track is not very wide, it has also attracted many companies to "compete".
"The New Year of Integrated Navigation and Parking is Coming, Supplier Analysis (I)" points out that the integration of navigation and parking is helpful to promote the coverage of intelligent driving in the whole scene, including high speed, low speed, driving and parking. Its biggest advantage is that it can significantly reduce the cost by reducing the number of controllers, at the same time, it can maximize the use of different sensors to increase the reuse of hardware, and maintain the flexibility of sensor configuration to obtain the best perception of the surrounding situation.
Yesterday’s article "The New Year of Integrated Shipping and Parking is Coming, Supplier Analysis (I)" took stock of some integrated shipping and parking suppliers, and today continues to introduce the characteristics of integrated shipping and parking schemes of other companies.
The DCU3.0 integrated domain controller scheme of Yingchi Technology is jointly developed with Horizon and NXP. It is based on NXP vehicle network processor S32G (which can cross the intelligent driving domain and intelligent gateway domain at the same time), Horizon high-performance car-level AI chip Journey 3 (which is responsible for the integrated parking and navigation scheme), and the high-performance computing software platform EMOS of Yingchi Technology.
DCU3.0 provides the greatest possible functional safety design, and can achieve greater flexibility and scalability for different high and low configurations of vehicles. The biggest challenge for most domain controllers in the market is the chip security level. Some chip manufacturers didn’t consider the functional safety requirements at the initial design, so they should meet the L3 functional safety requirements at a higher cost at the system level when composing L3 scheme. DCU3.0 not only provides the greatest possible functional safety design, but also considers the high and low configuration requirements of different models, which is both flexible and extensible, and meets the demand trend of the market in the integration of high-speed and low-speed intelligent driving ONE-BOX.
Neusoft Ruichi’s self-driving parking and parking integrated domain controller is designed independently, with TI TDA4. Through the integration of parking and driving functions and the sharing of sensors, L2+ enhances the perception ability, which improves the safety and user experience. This product is based on SOA software architecture, supports OTA upgrade through preset basic software and special middleware for autonomous driving, and provides rich "developer-friendly" tools for developers, which saves 20%-30% compared with the traditional 1V1R+APA technical scheme. Comply with information security standards and regulations, and functional security supports ASIL-B(D) level.
In terms of software, NeuSAR, a basic software product independently developed by Neusoft Ruichi and compatible with the latest standards of AUTOSAR, and related development tools are standard, and Neusoft Ruichi is equipped with an open SOA architecture for autonomous driving. The software architecture supports the requirements of L0-L4 autopilot function, and the software and hardware are decoupled efficiently through autopilot middleware and NeuSAR basic software, thus realizing the flexible deployment mechanism of reusable, configurable and extensible upper application. The product design is also equipped with special information security components and functional security mechanisms to support self-evolution and iteration of functions.
Based on the integrated platform of software and hardware, Fretek has built a multi-sensor integrated parking and traveling scheme, which can meet the product requirements of different levels of L2-L4 in the whole vehicle factory. The scheme is oriented to mass-produced ADAS and autopilot scenarios, including 1V, 1V1R and 1V3R;. The self-developed ADC domain controller provides high efficiency, flexibility and cost performance for the high-order autopilot solution of the software carrier platform, which not only supports the integration of parking and parking, but also supports the more powerful high-performance L2.9 driving function. Its integration scheme uses embedded hardware, continuous data optimization and OTA to realize continuous growth of functions and performance, more anthropomorphic system control, reduce driving fatigue, and make driving experience more comfortable and safer.
At the same time, the ADC20 domain controller platform, which integrates driving domain controller, parking domain controller and DMS controller with high-precision map and positioning, and rearview camera, can realize the functions of expressway navigation-assisted driving and AVP parking. It will be mass-produced and delivered to passenger car customers in the second half of 2022, and its functions can be continuously upgraded through OTA.
Moshi intelligent
Magic Vision Intelligence is one of the first autopilot companies to apply deep learning technology to ADAS. In order to enhance the recognition accuracy, such as vehicles recognizing objects or pedestrians from all angles directly behind and obliquely behind, Magic Vision Intelligence has also developed a deep learning framework Motovis UniVisity.
Magic Vision Intelligence has profound technical accumulation and experience in front-loading mass production projects in both driving and parking fields. For parking, Magic Vision Intelligent has developed an accurate and real-time environment perception and car body positioning system, launched a series of parking schemes from APA to AVP, and is accelerating the mass production of upgraded APA functions such as memory parking.
In terms of driving, Magic Vision’s early mass production projects were mainly L1 early warning, AEB (Automatic Brake System) /ACC (Adaptive Cruise) and other functions, and in 2018, it achieved mass production in BYD Song. Nowadays, the driving products of Magic Vision Intelligent have gone through three iterations. The third generation products are L2.5 class driving domain controller and parking integrated domain controller, which have NOA (pilot assisted driving) capability and will be produced in SOP (small batch) in the second half of this year.
As a full-stack ADAS and autopilot solution provider with algorithms, software and hardware, Magic Vision Intelligence has been deeply cultivated in four major areas: passenger car front-loading, commercial vehicle front-loading, active safety and vehicle-road coordination. It is estimated that the shipment volume will reach 400,000 units before 2022.
The intelligent pilot-assisted driving system products of Zhijia Technology are based on the deep learning perception algorithm, and the perception is integrated into the planning and control full-stack technology chain for self-research, which supports point-to-point one-click smooth operation. MAXIPILOT?2.0 can meet the L0-L2++ level function of automatic driving defined by SAE J3016?, and supports the NOM pilot assistance system and the integrated scheme of navigation and parking.
The system accesses high-precision maps, uses mapping and positioning technology, and makes full use of all kinds of information of the surrounding environment through FPP fusion path planning technology, and makes comprehensive judgment on information such as navigation, lane line, traffic flow, railings, guardrails and curbs with priority, which effectively reduces the takeover rate and improves the fluency of the system.
In order to improve the sensing accuracy and reduce the impact of accidental failure, the system adopts the dual redundancy design of target information; At the same time, it has the advantage of autonomous long-distance forward vision, which can help the parking system to achieve positioning and improve agility (parking at a faster speed). The system supports the deployment of OTA online upgrade technology and shadow mode, and provides users with a closed loop of data-driven technology, product and service innovation.
The application scenarios of MAXIPILOT?2.0 include passenger car application scenarios based on open roads and commercial vehicle application scenarios based on open roads.
The advantage of Yinjia Technology’s scheme lies in that it can integrate superior resources, build the full-stack software and hardware integration capability, and have the ability to decouple the software and hardware algorithms, and provide an all-round one-stop intelligent driving solution for the main engine factory, and quickly support it to meet the differentiated needs of different main engine factories.
In terms of the integration of navigation and parking, Yinjia Technology has five types of layouts, covering different intelligent driving stages from low speed to high speed and from assistance to automation. APA system identifies the available parking spaces around the vehicle through the combination of vision and radar, and controls the vehicle horizontally and vertically to complete parking after the driver selects the parking space, among which supporting the parking of mechanical parking spaces is the exclusive highlight function of Yinjia Technology; The HPA memory parking system can learn by itself, record the parking spaces in specific areas, and complete the functions of automatically finding and recording parking spaces within 100m, avoiding obstacles, automatically parking, and calling with one button. AVP can realize automatic driving in the whole process from the entrance of the parking lot to the parking space, while AVP+ aims at entering a strange parking lot without learning in advance, that is, it can complete the functions of autonomous library patrol, obstacle avoidance and remote summoning in the whole parking lot; In view of traffic congestion, TJP navigation system can avoid obstacles and change lanes; NOA automatic assisted navigation driving can realize point-to-point automatic driving according to high-precision positioning map.
Millimeter Zhixing is one of the fastest self-driving companies except Tesla. Less than half a year after the high-speed NOH landing, Mimo Zhixing took the lead in realizing the large-scale landing of urban assisted driving products.
The HPilot 1.0 system launched by Mimo Zhixing has the function of assisting driving in high-speed domain, and is mounted on five models of Great Wall Motor, namely Wei Mocha, Wei Latte DHT, Wei Macchiato DHT, Tank 300 and Haval Beast. In March, 2022, the brand-new tank 500 equipped with HPilot 2.0 system was officially launched, which added the integrated function of parking and navigation on the basis of supporting the assisted driving function in high-speed domain, further expanding the assisted driving range and experience of users. In April, HPilot 3.0 was officially released at 2022 HAOMO AI DAY, supported by the self-developed self-driving computing platform Magic Box 3.0, which supports the assisted driving function in urban areas. It will not only achieve landing during the year, but also further realize the boarding of more production models in the future.
According to statistics, from January to May, 2022, the average monthly growth rate of HM HPilot was over 200%. As of June, the total mileage of users equipped with HM Zhixing driver assistance system exceeded 10 million kilometers. Mimo Zhixing said that by the end of this year, the assisted driving system will be equipped with 34 different models of Great Wall Motor, and the total number will reach 1 million in the next two years.
Recently, Newtech’s intelligent driving solution based on the Orin chip platform in NVIDIA has been delivered in stages, and the mass production plan of large computing power domain has been further implemented. At this point, its high, medium and low computing power product solutions based on the automatic driving platform technology have all started substantial delivery.
New York Technology’s integrated intelligent driving project is the first domestic landing based on the standard version of Orin chip platform (110 TOPS), covering high-speed, urban and parking full-scene mass production applications. It is estimated that by 2023, the first model of the project will be delivered on the market, and the total mass production scale will be close to one million.
Under the premise of fully considering the functional safety requirements and various customization requirements, Newell Technology’s integrated intelligent driving solution can provide all mainstream functions such as navigation-assisted driving, congestion-following, parking assistance, conventional ADAS, etc. At the same time, it also reserves the upgrade space to support more innovative functions and the upgrade of automatic driving functions.
Its Orin platform scheme integrates a large number of cutting-edge technologies, and applies the data closed-loop tool chain with completely independent intellectual property rights for the first time to empower industrial customers. On the basis of conventional full-stack applications that provide perception, planning and control, Niuqi Technology helps to build an automated closed-loop data platform, drive the system to continue to grow itself, realize OTA upgrade in the whole life cycle, and meet the diverse needs of autonomous driving.
As a leading provider of autonomous driving technology in China, AutoBrain recently signed a strategic cooperation agreement with Hikang Automobile to jointly devote itself to the R&D and commercialization of mass-produced products in the field of intelligent driving, thus escorting intelligent travel. The two parties will cooperate in the development of integrated parking system, joint bid, joint promotion and other aspects, and establish a joint development team, including the systematic integration of the intelligent driving software system developed by AutoBrain with the high-performance and low-power intelligent driving domain controller, high-definition camera, ultrasonic radar and millimeter-wave radar of Haikang automobile, so as to provide automobile enterprises with mass production solutions including L1-L2.99 integrated parking.
AutoBrain has leading technology and development experience in the field of autonomous driving, providing high-performance front-loaded ADAS and mass-produced software products of autonomous driving system for passenger cars, covering all scene functions. The algorithm based on self-developed software can realize functions such as ADAS, NOP, CityPilot, etc. By establishing a data closed-loop cloud platform, the algorithm can be rapidly iterated and OTA upgraded, and through the fusion test verification scheme, the continuous iterative mass production system software can be provided for the main engine factory. As the earliest leader in the field of autonomous driving, AutoBrain has established strategic cooperative relations with OEMs such as BAIC Group and Sinotruk Group.
As a leading autonomous driving technology company in China, since the start of mass production, Immobile Technology has obtained the pre-loading mass production project in the domestic market and started to cooperate with international OEMs such as Japan and South Korea. Its full-stack self-developed intelligent driving system provides an integrated solution of driving, parking and parking based on SOA software architecture. The scheme integrates high-level ADAS functions such as traffic jam assistance (TJA), expressway assisted driving (NOA), automatic parking (APA), remote parking (RPA), memory parking (HPA) and parking service (AVP).
In order to help customers cope with the increasingly fierce market competition, in terms of hardware design, Immobile Technology has preset a memory space with high computing power and large storage, which can meet the needs of mass production and development of high-order autopilot functions and scenes in the next few years, and realize the continuous iterative upgrade and update of high computing power algorithms and software functions in the later period; In terms of software, the SOA-based software architecture presupposes basic software, standard middleware and various development tools. Deploying layered software architecture can realize efficient decoupling of software and hardware, make upper-level applications reusable, configurable, extensible, and portable on different platforms, ensure sufficient flexibility and openness, meet the individualized and differentiated development needs of different customers, significantly reduce development costs and shorten development cycles.
By integrating powerful visual perception algorithms, highly reliable application algorithms, efficient arbitration mechanisms, high-precision crowdsourcing and high-precision maps, Immobile Technology realizes driving and parking functions in different scenes, ensuring a comfortable, smooth and safe driving experience. The scheme also supports OTA unified upgrade and special information security components and functional security mechanisms. Rich interface protocols, high bandwidth and flexible communication mechanisms can meet the huge data requirements.
At present, the intelligent ADAS market is in a period of rapid growth, and the integrated solution of bank and berth will become the mainstream solution in the next few years, and mass production and loading will be ushered in in 2022 -2023. Integrating the driving and parking functions in the same domain control scheme will help to reduce the cost of intelligent driving, reduce the power consumption and weight of equipment, reduce the complexity of the whole vehicle, and accelerate the mass production process of the integrated driving and parking scheme.