「Qingzhen」intelligent networking
The company has made attempts to develop original automotive technology. It develops prospective and generic platforms and core products and technologies of new energy and intelligent networking automobiles. It establishes a new intelligent networking auto ecology in areas such as vehicle operating systems, intelligent cockpits, visual perception algorithms, digital maps, human-vehicle road cloud information aggregation platform, 4D mm-wave radar, information security, vehicle-level SiC power module, cloud service platform, etc., to enhance travel and improve life quality.
Vehicle operating system
The China Association of Automobile Manufacturers awarded the company the "2021 China Automobile Supply Chain Excellent Innovation Achievement"
Comprehensive customization based on international standards, as well as concurrent creation of MOS, CP, and CAP products
A complete toolchain includes design, configuration, development, and deployment functions
MOS product has passed ASILD certification, the highest function safety certification of ISO26262
MOS and CAP-AD products have passed ASPICE CL3 certification
Smart cockpit technology base

Standard version platform for cockpit area control

Low latency: Built-in R core and VDSP engine, which support fast start
Multi-scheme: Support virtualization and hard isolation
High coupling: P2P compatibility of upgraded chip and standard version
High standard: Meet the vehicle grade test standard
High safety: Independent safety island and high-level function safety

High level version platform for cockpit area control

Large hash rate: CPU: 93K, NPU:8T
Multi-interface: Support 1-machine and 6-screens output, 12-camera input
Low power consumption: 6nm, which reduces lower consumption by 18% than 7nm
Low cost: Integrated 5G baseband, with a low system cost; smartphone chip modification, which contributes to a lower cost
High standard: Chipset meets the AEC-Q100 vehicle grade test standard; module meets AEC-Q104

Core technology platform

Software platform: Layered decoupling, platform reuse, differential configuration, continuous evolution, and rapid realization
AI platform: Facial features, gesture features, and fusion perception
Product/Design platform: User research, scenario definition, and design innovation

AI engine & vision solutions
The front-view algorithm may be able to meet the requirements for mass production
Front-view and surround-view sensing algorithms developed in-house could achieve end-to-end output via a two-stage convolutional neural network model (NNM), meeting the sensing requirements of ADAS, HWP, and AVP
Self-developed AI engine covers four platforms
The self-developed AI engine has four platforms: data collecting, annotation, model training, and algorithm assessment, and it supports 15 annotation tools, including 2D and 3D annotation
Wider range of application
Create a traffic scene database with Chinese characteristics, and support full scenario coverage
Wider range of goals
The self-developed NN algorithm can detect 8+ objects, recognize 6 types of secondary attributes and perform 16+ semantic segmentations
Improved perception
Average recognition rate is>98%, and separation MIoU is>90% under standard conditions
Visual perception algorithm
AI engine
Intelligent Vehicle Cyber Range
First intelligent vehicle cyber range of automotive industry in china
Undertake cyber security test projects of leading enterprises in the automobile industry, and provide testing and verification services for parts、vehicle、and network applications, and event services.
Services of "I-VCR Intelligent Vehicle Information Security Evaluation Index"
Provide unified intelligent networking vehicle cyber security evaluation services
Automotive SiC power module
Higher operating voltage
VDSS=1,200V, which can better meet the needs of an 800V automotive voltage platform
Tj: 175℃ at maximum
'-40℃~175℃, which could reduce heatsink volume and dissipation cost
Higher operating frequency
Fop≥20KHz, reducing the volume of peripheral components and contributing to a compact structure, and lower costs
Lower power loss
RDS(on)<2.5mΩ, which could improve energy efficiency and cruising range
Intelligent driving system
Leading performance
The sensing algorithm has a 97% accuracy/recall rate, the lateral and longitudinal control precision of parking and driving is less than 10cm, and the fusion positioning accuracy is less than 10cm
High-efficiency multiplex
Self-developed unified platform algorithm software architecture that enables multiplex and is equipped with many platforms and processing units with strong customization ability
Various scenarios
Unmanned tasks covering points A-B in cities and high-speed scenarios, as well as autonomous parking tasks in parking lots; support collaborative sensing and regulation technologies; and can achieve application in restricted spaces such as mines and logistics
Iterative intelligence
To improve algorithm development efficiency, we implement closed-loop features such as scene-to-map and twin modeling, scene-to-virtual simulation, acquisition and annotation training, iterative toolchain, etc.
V2X protocol stack and 5G V-Box vehicle-mounted intelligent terminal

Passed the conformance test of the China Academy of Information and Communications Technology, the V2X "Four Spans" (H.S.X) pilot application certification, and the 2022 IoV C-V2X Security Verification Certification

V2X Protocol stack
Cover all national standard DAYI and DAYII application scenarios, and support scene fusion of V2X, L3/L4 intelligent driving system, and intelligent cockpit system
Loose coupling, easy adaptation, and interface expansion; enable the rapid development of bespoke application scenarios
It may ensure packet loss rate and delay can meet communication needs under large-scale scenarios using a multi-thread + queue cache method
Support for quick module deployment, as well as demand-based ordering and use of various modules
5G V-Box vehicle intelligent terminal
The main chipset is entirely homemade, allowing for V2X, T-BOX, and central gateway integration
Cutting into modules facilitate integration into the cockpit or the central computing domain controller, allowing for rapid customized development.
Intelligent transportation solution for vehicle-road-cloud integration
Intelligent road building
Create intelligent roadways that match various scenarios by providing point-level surveys and tailored designs
Construction of the checheng.com platform
Create three sectors: "cloud driving application platform, traffic management platform, and urban management platform" to fully integrate smart city, intelligent transportation, and intelligent networking vehicle
New network construction
Create a network that meets autonomous driving and city car application needs, and achieves full 5G coverage throughout the region
Build application scenarios
Carry out customized, large-scale, and sustainable unmanned minibus and Robo-taxi operating modes, provide intelligent travel experiences with high S & T connotation, and enable urban travel services
Formulation of measures and codes
Formulate intelligent networking vehicle-related policies and regulations, and help establish and research local road traffic management laws and regulations
4D-mm wave radar
Wide detection range
4D front radar: Ranging range 0.5m~ 260m, azimuth -60°~+60°, angle of elevation -12°~+12°
4D angle radar: Ranging range 0.25m~120m, azimuth -75°~+75°, angle of elevation -15°~+15°
High sensing accuracy
4D front radar: Ranging accuracy ±0.12m, speed measurement accuracy ±0.1m/s, angle measurement accuracy ±0.3°
4D angle radar: Ranging accuracy ±0.13m, speed measurement accuracy ±0.1m/s, angle measurement accuracy ±0.3°
Strong perception ability
4D front radar: Range resolution 0.24m, velocity resolution 0.13m/s, angle resolution 2.5°
4D angle radar: Range resolution 0.25m, velocity resolution 0.13m/s, angle resolution 2.5°
Multiple outputs
Support two output modes: track and point cloud; a maximum output of 64 track targets and 256-point clouds

Multi-target detection
tracking

Point cloud

Prefusion
perception

Omni-bearing
detection

Centimeter-level
accuracy

Application in different
scenarios

All-weather
application

Light
weight

4D forward mm-wave radar

4D angular mm-wave radar

Vehicle intrusion detection and prevention system
ICCE and AUTOSEMO awarded this case the "2022 AUTOSEMO Innovative Product Excellent Case" designation
Microservices and high concurrency (VSOC Vehicle Security Operations Center)
Leading lightweight system architecture design, million-level concurrency, elastic extension, and vehicle operation management that allows attack event situation display, big data analysis, and automated rule orchestration
Self-adaptive and high-performance (CIDS intrusion detection)
Self-adaptive AUTOSAR; automatic rule generation support, and realize vehicle control intrusion detection of function safety linkage
Full coverage and vehicle grade (HIDPS+NIDPS)
High reliability, lightweight, and multiple safety designs; meet cockpit and autopilot real-time protection needs in complex scenarios
Compliant, accurate, and fast (DS data sentinel)
Accurate compliance detection; support extension of detection rules; meet cockpit and autopilot data security and compliance requirements
Basic platform for dynamic high-precision map
Create a national-level basic platform for intelligent driving vehicles with dynamic high-precision mapping
T3 share of Dongfeng Motor, China Automotive Innovation Corporation, Nanjing LingHang Technology Co., Ltd. (T3 travel)
Create a Level 10 million intelligent driving automobile dynamic sensor data ecosystem
Collect millions of passenger vehicle dynamic sensor data, millions of commercial vehicle dynamic sensor data, and millions of travel service vehicle dynamic sensor data
Dynamic, automated, high-precision map toolchain with entirely independent IPRs
Vastness · Data annotation and algorithm development platform, Sailing · data aggregation and processing platform, Voyage · map construction and dynamic update platform, Leading · map compilation and service release platform
Ensure the safety and compliance of intelligent driving vehicle dynamic data over the whole life cycle
Support relevant authorities' safety oversight, as well as the safety and compliance of national geographic information and personal private information
Total factor and full scenario-based dynamic high-precision map service
High-precision maps with absolute accuracy of 20 cm, complete scenarios of expressways, regular roads, and parking lots, as well as road, lane, and object models with 100+ elements
Intelligent networking cloud control platform

Vehicle to Infrastructure Connectivity

Connect and gather the vehicle and road-end equipment operation data to accelerate the development of autonomous driving and IVIC vehicles

Vehicle Management

Support the operation and management of autonomous driving vehicles hydrogen-powered vehicles, and vehicle-road system vehicles

Demonstration Zone operation

Create a typical twin mapping scenario in the demonstration zone and empower the monitoring and command center of intelligent park by using technologies such as high-precision maps artificial intelligence (AI), big data, etc.

Flexible and scalable

Flexible system deployment support for the expansion of various vehicle types, roadside equipment, and third-party platforms

Data service, privacy security circulation, and identification resolution node
XML 地图