Vertical evolution and horizontal comparison of NIO, Xpeng and Ideal autonomous driving capabilities

Latest update time：2021-06-18

Reads：

introduction:

With the arrival of the "intelligent" wave, major domestic and foreign car companies have begun an arms race in the field of autonomous driving; new domestic car-making forces such as Weilai, Xiaopeng, Ideal, etc. are gearing up and showing off their magical powers, just to get there You can get an entry ticket to the autonomous driving track before 2025.

01

Vertical evolution path of autonomous driving system

1. NIO-Autonomous Driving System Vertical Evolution Path

NIO-Autonomous Driving System Vertical Evolution Path

In December 2017, the driving assistance system NIO Pilot was released for the first time on NIO Day;
In June 2019, the vehicle-machine system NIO OS was upgraded to NIO OS version 2.0.0 through OTA. The NIO Pilot system completed its first major upgrade, adding 7 new functions, including high-speed automatic assisted driving, congestion automatic assisted driving, and turn signal control. Lane changing, road traffic sign recognition, lane keeping function, front traffic warning, automatic parking assist system;
In February 2020, the vehicle-machine system NIO OS was upgraded to version 2.5.0. The NIO Pilot system added automatic emergency braking (pedestrians and bicycles), overtaking assistance, and lane avoidance functions;
In October 2020, the vehicle-machine system NIO OS was upgraded to version 2.7.0. The NIO Pilot system added the use of high-precision maps and added an automatic assisted navigation driving function (NOA); on the basis of the rear crossing vehicle warning function , the active braking function has been added; the driver fatigue monitoring function has been upgraded, and a new camera has been added to identify the driver's facial, eye and head posture characteristics, and comprehensive judgment can be made through a variety of information;
In January 2021, the vehicle-machine system NIO OS was upgraded to version 2.9.0. The NIO Pilot system added visual fusion automatic parking and vehicle short-distance summoning; the NOA function was optimized: the function was enhanced in active lane changing and merging. Stability in the scenario of entering/leaving the main road;
In January 2021, the autonomous driving system NAD was released on "NIO Day", including the NIO Aquila NIO super-sensing system and the NIO Adam NIO supercomputing platform; and will gradually achieve coverage of highways, urban areas, parking, etc. Point-to-point autonomous driving function experience in the scene.

2. Xpeng-Autonomous Driving System Vertical Evolution Path

Xpeng-Autonomous Driving System Vertical Evolution Path

1) Xpilot2.0 : the first official product implementation version of the autonomous driving assistance system Xpilot

In January 2019, the OTA upgrade was pushed to Xpeng G3 users for the first time. The upgraded Xpilot2.0 system added a vehicle key summoning function; the automatic parking function was optimized: automatic parking will adapt to more scenes and support shadows. parking spaces, ground lock identification is more accurate;
In June 2019, Xmart OS version 1.4 was pushed to Xpeng G3 users via OTA. After the upgrade, Xpeng G3 will have the ICA intelligent cruise assist function (LCC lane centering assist above 60km/h);

2) Xpilot2.5 : L2 driving assistance system achieves mass production application

In July 2019, Xmart OS version 1.5 was pushed to G3 users through OTA. After the upgrade, the Xpeng G3 will have driving assistance functions such as automatic lane change assistance and traffic jam assistance;

3) Xpilot3.0 : Release the high-speed scene NGP-automatic navigation assisted driving function

Second quarter of 2020: XPILOT 3.0 hardware system will be implemented on Xpeng P7;
Fourth quarter of 2020: The basic functions of XPILOT 3.0 will be delivered, that is, ACC/LCC/ALC and other functions will be opened through OTA;
First quarter of 2021: Xmart OS version 2.5.0 was pushed to P7 users through OTA. Through this upgrade, the Xpeng P7 equipped with the XPILOT 3.0 system has the highway-NGP automatic navigation assisted driving function;
June 2021: The public beta of Xmart OS version 2.6.0 started. Through this upgrade, the Xpilot3.0 system added parking lot memory parking, smart high beams and driver status monitoring; at the same time, it also supports NGP automatic navigation assisted driving. The function and lane centering function have been optimized;

4) Xpilot 3.5: Release the NGP-automatic navigation assisted driving function in urban scenes

Xpilot 3.5 adds the NGP function in urban scenes and optimizes the parking memory parking function.
It is planned to start delivering Xpeng P5 to users in the fourth quarter of 2021. The urban scene NGP function in the Xpilot3.5 system will begin internal user testing at the end of the year and will be available to users through OTA in early 2022;

3. Ideal - vertical evolution path of autonomous driving system

Ideal vehicle model development and vertical evolution path of autonomous driving systems

Li Auto's autonomous driving research and development progress is slightly behind NIO and Xpeng. However, since the company's IPO in the United States in July 2020 and the former Visteon chief architect Wang Kai joining Li Auto in September, Li Auto's progress has obviously Accelerating the pace of research and development in the field of autonomous driving;

Li Xiang announced Li Auto’s roadmap for autonomous driving development in an interview in 2020:

1) 2021-2022 : Realize automatic navigation assisted driving function;

2) 2023 : Launch of a new model, Li X01 - equipped with a hardware system supporting L4 autonomous driving capabilities

3) 2024 : Achieve L4 autonomous driving capabilities through OTA on mass-produced models

02

Autonomous Driving System - Perception Solution

Appendix 1: Comparison of sensing system solutions among three car manufacturers

Note: √ There is this configuration, × There is no such configuration

1. NIO-autonomous driving system perception solution

Comparison of NIO Pilot and NAD system perception solutions

Note: 1) √With this configuration, N/A Not applicable, — unknown

2) Some of the 11 8-megapixel high-definition cameras of NIO ET7 are supplied by Lianchuang Electronics; by querying the official website of Lianchuang Electronics, we learned that the company’s main application areas for vehicle cameras are surround view and rear view, so it can be It is inferred that the surround view camera of NIO ET7 is most likely supplied by Lianchuang Electronics.

Compared with the NIO pilot system, the NAD system's perception solution is more complete:

摄像头方面 -增加了侧视和后视ADS摄像头；但是ADS前视却由3个摄像头变成了2个摄像头；

注：原因推测 -NAD系统采用ADS前视摄像头*2+侧前视摄像头（瞭望塔式布局）*2+前置激光雷达*1，完全能够覆盖到NIOpilot系统采用三目摄像头情况下的视野探测范围；
激光雷达 -增加了一个前置激光雷达
新增 C-V2X 通讯模块；

1 ）NIO Pilot- 感知系统

感知传感器 ：ADS摄像头*3+环视摄像头*4+毫米波雷达*5+超声波雷达*12+DMS摄像头*1

NIO Pilot 传感器配置（图片来源：蔚来官网）

重点传感器介绍：

a. 前视摄像头： 三目摄像头（180万像素，蔚来自主研发）

52°中距离摄像头：主要用于一般性道路状况监测
28°长焦远距离摄像头：探测远距离目标和红绿灯
150°广视角近距离摄像头：探测车身侧面和短距离插队的车辆

三目摄像头主要实现的功能如下图所示：

实现功能 ：道路标识识别、远近光自动控制、车道偏离预警、车道自动保持、紧急制动辅助、前向碰撞预警

三目摄像头功能实现（图片来源–蔚来官网）

b． 毫米波雷达 ：1个前置长距离雷达 + 4个角雷达

实现功能 ：车道变换预警、车辆盲点监控、侧方开门预警、前向来车预警、后向来车预警

毫米波雷达功能实现（图片来源–蔚来官网）

2 ） NAD- 感知系统

蔚来Aquila超感系统： 感知传感器 + 高精度定位单元（GPS+IMU）+车路协同感知V2X

感知传感器： 激光雷达*1 + ADS摄像头*7 + 环视摄像头*4 + 毫米波雷达*5+超声波雷达*12

蔚来超感系统（图片来源–蔚来官网）

重点传感器介绍：

a 、激光雷达（图达通- Falcon）性能参数：

类型： 双轴旋转镜扫描式混合固态激光雷达
探测距离： 250m@10%反射率
FOV （H&V）： 120°*30°
角分辨率（H&V）： 0.06°*0.06°
激光波长： 1550nm

b ．ADS高清摄像头：

7 颗ADS摄像头（8MP）： 前视*2（前风挡）+ 侧前视*2（车顶前部两侧）+ 侧后视*2（翼子板） + 后视*1（车顶后部居中）

ADS 摄像头布置位置示意图

侧前视摄像头-瞭望塔布局： 传感器视线可有效越过遮挡，减小盲区

减小盲区： 在城市场景下，传感器视线很容易被绿化带和车辆遮挡，相较于安装在 B 柱和后视镜的摄像头，部署在车顶的高位侧前摄像头，能够减小盲区；
作为前向视觉冗余： 布置在车顶高位侧前摄像头，因位置高，故视野广，提高了前向视觉的冗余度。即使前向主摄像头不工作，依靠两个高位侧前摄像头，仍然可以实现前向视觉的完整感知。

蔚来ET7侧前视摄像头（图片来源-蔚来官网）

2. 小鹏 - 自动驾驶系统感知解决方案

不同版本Xpilot 系统感知方案对比

注： √有此配置，N/A 不适用，— unknown

Xpilot 3.0 相比于Xpilot2.5 ：

ADS 摄像头 ：前视摄像头由单目升级成三目摄像头+1颗前置单目摄像头，增加侧视（4颗）和后视（1颗）共5颗摄像头
增加了1颗驾驶员监控摄像头
毫米波雷达 ：新增了两颗后置角雷达
新增了 高精地图 的应用

Xpilot 3.5 相比于Xpilot3.0：

ADS 摄像头： 前视中去掉了1颗前置单目摄像头

注：原因推测 - Xpilot3.5 系统中在车辆前部布置有两颗激光雷达，完全可替代之前的单目摄像头，起到感知冗余的作用。
激光雷达： 新增了2颗前置激光雷达

1 ）Xpilot 2.5 - 感知系统

感知系统配置 ：ADS摄像头*1+环视摄像头*4+毫米波雷达*3+超声波雷达*12

Xpilot 2.5 传感器配置（图片来源：网络）

注：紫色圆圈-前置中距毫米波雷达*1，橙色圆圈-后置近距毫米波雷达*1，绿色圆圈-环视摄像头*4，蓝色圆圈-前置主摄像头*1，红色圆圈-超声波雷达*12。

重点传感器介绍：

a. 前置毫米波雷达（博世-MRR Evo14）

中距离雷达 ：拥有广角/窄角两种探测角度，探测距离分别为：100米和160米；布置于前保险杠中间位置，主要用于前方目标物的探测与追踪。

b． 后置毫米波雷达（博世）

近距离雷达 ：探测距离0.36~80m；布置于后保险杠两侧，主要用于车身两侧后方障碍物的探测

2 ） Xpilot 3.0 - 感知系统

感知系统配置： 感知传感器 + 高精度地图（高德）+高精度定位

感知传感器： ADS摄像头*9+环视摄像头*4+毫米波雷达*5+超声波雷达*12+DMS摄像头*1

重点传感器介绍：

9 颗ADS摄像头 ：前置三目（前风挡）+前置单目（前风挡）+侧前视*2 （后视镜底座） +侧后视*2 （翼子板） +后视*1 （牌照板上部）

a. 前置三目摄像头：

长距离窄视角摄像头（28°） ：主要用于前向移动物体探测，如自动紧急制动，自适应巡航，前向碰撞预警

性能参数 ：探测距离150m以上，分辨率-1828*948，像素-2MP，帧率-15fps
中距离主摄像头（52°）： 主要用于交通信号灯识别、车道识别以及前方移动物体的探测，如交通信号灯探测，自动紧急制动，自适应巡航，前向碰撞预警，车道偏离预警

性能参数： 像素-2MP，帧率-60fps
短距离广视角摄像头（100°）： 主要用于防加塞、雨量探测以及交通信号灯识别

性能参数： 像素-2MP，帧率-60fps

b. 前置单目-安全冗余感知摄像头（前风挡）： 像素-2MP，帧率-69fps

作用： 作为前置三目摄像头感知失效情况下的感知冗余传感器

3 ） Xpilot 3.5 - 感知系统

感知系统配置： 感知传感器 + 高精度地图（高德）+高精度定位

感知传感器： 激光雷达*2 + ADS摄像头*8+环视摄像头*4+毫米波雷达*5+超声波雷达*12+DMS摄像头*1

重点传感器介绍

a ．激光雷达（大疆Livox-HAP）性能参数：

类型： 双棱镜扫描式混合固态激光雷达
探测距离： 150m@10%反射率）
FOV （H&V）： 120°*30°
角分辨率（H&V）： 0.16°*0.2°
激光波长： 905nm

b ．8颗ADS摄像头： 前视*3（前风挡，2MP像素）+ 侧前视*2（后视镜底座）+ 侧前视*2（翼子板）+ 后视*1（牌照板上部）

小鹏P5 – ADS摄像头布置位置示意图

3. 理想-自动驾驶系统感知解决方案

理想汽车自动驾驶系统感知方案对比

注： √有此配置，N/A 不适用，— unknown,（E）据推测

2021款理想One相比2020款理想One的感知系统：

a． 前置摄像头性能参数提升 ：老款理想One前置单目摄像头为130万像素，水平视角52°；新款理想One配置的单目摄像头为800万像素，水平视角120°

b．增加了 毫米波雷达 数量：新增加了4个角雷达

1 ）2020款理想One - 感知系统

传感器配置 ：ADS摄像头*1+环视摄像头*4+毫米波雷达*1+超声波雷达*12

理想one 传感器布置示意图

注：上图可以看到前风挡位置有3颗摄像头：a. 上方的2颗中只有②为ADS摄像头，即用于行驶过程中的驾驶辅助摄像头；b. 另外1颗摄像头①为数据采集与收集摄像头，专门用于收集道路信息和驾驶场景等数据信息，便于后续对相关算法的训练与优化；c. 最下方的③摄像头为 行车记录仪摄像头 ；

重点传感器介绍：

前置单目ADS摄像头（豪威科技-OV10642）和环视摄像头（豪威科技-OV10640）

OV10642 与OV10640性能参数对比

2 ）2021款理想One – 感知系统

传感器配置 ：ADS摄像头*1+环视摄像头*4+毫米波雷达*5+超声波雷达*12

重点传感器介绍：

a. 前置单目摄像头： 8MP像素，水平视角120°，探测距离200m

b. 毫米波角雷达： 博世第五代雷达，水平视角120°，探测距离110m

3 ）理想X01 - 感知系统

感知系统配置： 激光雷达*1 + ADS摄像头*8（E）+环视摄像头*4+毫米波雷达*5（E）+超声波雷达*12+DMS摄像头*1

03

自动驾驶系统 - 计算平台

附表2. 三家造车新势力自动驾驶系统计算平台对比

1.蔚来 – 自动驾驶系统计算平台

蔚来自动驾驶计算平台信息梳理

1 ）NIO Pilot 自动驾驶计算平台

硬件：Mobileye- EyeQ4芯片（感知处理运算）+恩智浦-S32V芯片（决策处理运算）

a. EyeQ4 性能参数 ：算力 -2.5 TOPS，功耗 -3W， 响应时间 -20ms， 制程工艺 -28nm， 芯片类型- ASIC

b. S32V 性能参数：算力 -9200 DMIPS， 处理器 -4核64位CPU Arm®Cortex®-A53 1GHZ ，4核32位CPUArm®Cortex®-M4 133MHZ

S32V 结构框图（图片来源-NXP官网）

2 ）NAD 自动驾驶计算平台

蔚来超算平台NIO Adam（由4颗英伟达Orin芯片构成）性能：

蔚来超算平台-NIO Adam（图片来源：蔚来官网）

4* 高性能专用芯片： 2颗主控芯片+1颗冗余备份芯片+1颗群体智能与个性训练专用芯片

—— 2主控芯片 ：实现NAD算法的全栈运算，包含多方案相互校验感知，多源的高精度定位，多模态的预测与决策；充足的算力确保NAD系统能够游刃有余地去处理复杂交通场景；

—— 1冗余备份芯片 ：任何一个主芯片失效，NAD都可以确保安全

—— 1群体智能与个性训练专用芯片 ：可以加快NAD的进化进度，同时又可以针对每个用户的用车环境进行个性化的本地训练，提升每位用户的自动驾驶体验；
超级图像处理流水线 ：超高带宽的图像接口，ISP每秒可处理64亿像素
超高骨干数据网络 ：将所有传感器和车辆系统的信号输入实时无损地分配到每一算力核心

2.小鹏–自动驾驶系统计算平台

Xpilot 系统计算平台信息梳理

1 ）XPilot 2.5 自动驾驶计算平台

硬件： 采用 Mobileye的EyeQ4 模块

EyeQ4 芯片：算力-2.5 TOPS，功耗-3W，响应时间-20ms，制程工艺-28nm 芯片类型-ASIC

2 ）Xpilot 3.0自动驾驶计算平台

Xpilot3.0 计算平台： 英伟达Xavier SoC+ 英飞凌Aurix MCU

a． 英伟达Xavier SoC ：算力-30TOPS，功耗-30W，安全等级-ASIL D，视频处理器-8K HDR，CPU-定制8核ARM架构，GPU-512核VoltaGPU

b． 英飞凌Aurix MCU： （RISC处理器核心、微控制器和DSP整合在一个MCU）

算力-1.7DMIPS/MHz，时钟频率-单核300MHz，制程-40nm，安全等级-ASIL D

通过下面的Xpilot3.0系统的软件堆栈可以看出：

自研部分包括 ：XPU的平台软件（Xpilot系统架构）以及Xpilot 系统应用层软件：包括OTA、诊断服务、虚拟化、摄像头服务、CAN服务、自动驾驶应用（感知、定位、预测、路径规划等）等

Xpilot 3.0系统软件堆栈（图片来源–小鹏）

3.理想–自动驾驶系统计算平台

理想自动驾驶计算平台信息梳理

1 ）2020款理想One搭载的自动驾驶计算平台

图像感知处理模块 ：采用Mobileye EyeQ4芯片
控制决策模块： 由易航智能研发

2 ）2021款理想One搭载的自动驾驶计算平台

处理芯片 ：地平线征程3*2

征程3基础参数信息梳理

3 ）理想X01搭载的自动驾驶计算平台

硬件：

自动驾驶域控制器： 德赛西威
主控芯片：SP控制器芯片- 英伟达Orin-x

操作系统： 理想X01将搭载其自研的实时操作系统Li OS

Li OS基于Linux内核深度定制开发，定制开发的部分包括：a. 内核的文件系统、IO系统和Boot引导等核心部分; b. 与其它应用层通讯的中间件部分；应用层部分则选择与战略供应商合作开发；

04

自动驾驶系统 – 功能实现

附表3：三家造车新势力智能驾驶系统付费模式对比

附表4：三家造车新势力自动驾驶系统实现功能对比

注： √ 有此配置，×无此配置

1.蔚来–自动驾驶系统功能实现

NIO Pilot 和NAD所能实现的功能

1）NIO Pilot系统基础功能免费，对于精选包和全配包采用一次性付费模式；

2）NAD采用按月订阅付费的模式：除一些基础功能外，如果想要拥有NAD的完整功能需采用“按月开通、按月付费”的服务订阅模式，即ADaaS（AD as a Service）。

3） NAD系统会在NIO Pilot系统的基础上，新增一些基于场景的、集成式的高阶自动驾驶类功能：换电站自动泊车，NAD低速及泊车自动驾驶等功能

2. 小鹏–自动驾驶系统功能实现

Xpilot系统实现的功能

注：带“ * ”的功能，属于选配装置，后期订阅付费后可通过OTA升级获取

1 ）Xpilot 3.0相比Xpilot 2.5：

Xpilot 3.0具备较强软件升级能力，后续通过OTA升级能力可新增： 智能远光灯、停车场记忆泊车、NGP自动导航辅助驾驶（高速路）、SR自动驾驶环境模拟显示

2 ）Xpilot 3.5相比Xpilot 3.0：

稳定性提升：可更好地解决黑夜、逆光、弱光、隧道明暗交替等疑难场景；
新增 城市场景NGP -自动导航辅助驾驶功能
提升停车 记忆泊车功能 体验：在停车场更好应对出入库车辆、对向来车、行人穿梭、连续直角弯等复杂场景

3. 理想–自动驾驶系统功能实现

理想-自动驾驶系统实现的功能对比

1 ）2020款理想One ：主要是实现L2级驾驶辅助功能，在L2+高级别驾驶辅助功能上有所欠缺；

a. 由于车辆没有配置能够监控车辆后方的长距离传感器，比如说毫米波雷达或侧后视/后视ADS摄像头，因此在快速行驶时车辆后方防护能力较弱；较难以较好地实现车辆后方碰撞预警、并线辅助以及自动变道等功能；

b. 前向传感器的感知能力具有一定的局限性

在2020年理想One接连发生追尾大货车的事故，这些事故原因主要有以下两种情况：第一：理想One当时开启了辅助驾驶功能，但前车侵入本车道1/3-1/2时，ADAS仍未识别前车；第二：理想One的辅助驾驶感知传感器对侧前方车辆识别能力较差，对于在光线弱的黑夜更是形同虚设。

因此，除了软件算法因素之外，传感器前向感知能力的局限性也是另外一个重大原因，首先2020款理想One的前置摄像头是单目，且视场角为52°，并且车辆两侧也没有布置角雷达，对于光线较弱的情况下感知能力较差，因此对于侧面快速插入的车辆识别较晚。

因此2021款理想One自动驾驶系统相关硬件配置的升级也自然是情理之中的事情了。

2 ）2021款理想One ：前置摄像头采用800万的高清摄像头，水平视角120°，并且增加了4个角雷达，增强了对后方以及前方两侧中长距离目标物体的感知能力；相比老款理想One，增加了红绿灯识别功能、前方车辆横穿提醒、后方车辆横穿提醒、开门防撞提醒等功能；

05

结语

1. Judging from the autonomous driving R&D plans and self-research layouts of the three new car manufacturers, their strategic layout directions are relatively consistent, and they all want to build full-stack self-developed autonomous driving capabilities; because the hardware of future models will converge , the enterprise's software capabilities or software and hardware integration capabilities will be its core barrier;

2. In terms of sensing solutions for future high-end autonomous driving, the three new car-making forces will all adopt the "strong sensing" route of redundant sensing of multiple sensors including lidar + millimeter wave radar + camera; however, the evolution of autonomous driving technology will In addition to the combination of software and hardware, data is the key. Only by continuously accumulating local user data under China's complex road conditions and continuously optimizing software algorithms can we create a differentiated autonomous driving experience; Chinese local companies have a better understanding of road scenes with Chinese characteristics. Due to the driving habits of Chinese users, there is an opportunity to iterate on self-driving technology with a better experience in Chinese scenarios; for example, some special functions suitable for Chinese roads: traffic jam warning, night overtaking reminder, large truck avoidance, etc.;

3. In terms of autonomous driving computing platforms, NIO, Xpeng, and Ideal all chose the relatively closed Mobileye solution in the early days, which made it difficult for them to effectively obtain and utilize underlying visual data. Therefore, they later turned to more selective and open solutions. Cooperate with NVIDIA; in the future, powerful companies will form a research and development model of self-research of AI chips + self-research of autonomous driving domain controllers (cooperative development). This model, combined with self-developed software algorithms, can give full play to the capabilities of the autonomous driving computing platform optimal performance;

4. In terms of functional implementation of mass-produced models, more attention will be paid to the implementation of specific user scenarios; starting from user scenarios, we will gradually realize point-to-point autonomous driving in three major scenarios: parking, highways, and urban areas;

5. Currently, autonomous driving technology is in a stage of continuous iterative development. OTA can realize rapid iteration of vehicle software, and can continue to add new functions or optimize existing functions to vehicles that have been delivered to users; however, OTA for new functions is also stronger than the hardware. Relatedly, if the hardware does not support it, software alone cannot achieve it. However, most of the hardware installed on current mass-produced models cannot be completed in one step. OEMs can only do their best based on their respective strengths - "Be forward compatible and try to be forward compatible."

— END —

Submission cooperation: 18918250345 (WeChat)

References:

1. NIO analysis of NIO’s new autonomous driving system

https://mp.weixin.qq.com/s/NKc0cP6Fc_xRlp-ui4lc7A

2. NIOPilot pilot assistance: How to define the world’s second best?

https://www.bilibili.com/read/cv7878209/

3. More stable than Tesla, be the first to experience Weilai NOP pilot assisted driving

https://www.sohu.com/a/423524857_118021

4. NIO announced the progress of autonomous driving: abandoning the pure visual route and doing multi-sensor fusion

https://www.sohu.com/a/301101089_256868

5. The road to upgrade NIO’s automatic assisted driving Pilot (with detailed explanation of super full functions)

https://mp.weixin.qq.com/s/QiS0e6Yye54IIYRtvi_C-A

6. From the 2019 PPT, explore the autonomous driving of Xpeng P7 (XPilot3.0)

https://zhuanlan.zhihu.com/p/148570354

7. From G3 to P7, 2.5 to 3.0, the evolution history of Xpeng XPILOT advanced assisted driving system

http://www.evask360.com/article-996.html

8. Vice President of Xpeng Autonomous Driving: Autonomous driving, system architecture pilot function evolution and P7 case analysis

https://www.sohu.com/a/355060121_99957909

9. ICA Intelligent Cruise Assist is launched and Xpeng G3 intelligent driving enters the L2 era

https://www.sohu.com/a/320206563_117727

10. From G3 to P7, 2.5 to 3.0, the evolution history of Xpeng XPILOT advanced assisted driving system

https://mp.weixin.qq.com/s/tOZouR8t_-Ixlg4CN9y3TA

11.A picture overview of Xpeng P7 suppliers

https://mp.weixin.qq.com/s/jG7sz8GGgtaXnzdE4h6Lyw

12. Core information of Xpeng P7 OTA upgrade briefing session

https://zhuanlan.zhihu.com/p/347364948

13. The success rate is 99.99%. Xiaopeng disclosed key OTA data and will soon launch new features that will surprise you.

https://mp.weixin.qq.com/s/SW3qXhVWJnNG0kJqhl0b9w

14. From Nokia to new cars, a chief architect decided to devote himself to "ideal"

https://zhuanlan.zhihu.com/p/254823116

15. Dialogue with Ideal CTO Wang Kai: Decrypting the Admission Ticket for 2025 Autonomous Driving Enterprises

https://zhuanlan.zhihu.com/p/256204278

16. OmniVision Group's comprehensive vehicle vision solution equips the ideal ONE car with the "smart eye" for intelligent driving

https://mp.weixin.qq.com/s/TRNG0UOhDJywQbUZo9rKCA

17. What do you think of the Ideal ONE autonomous driving system?

https://www.sohu.com/a/337831069_607980

18. Li Xiang announced Li Auto’s autonomous driving roadmap for the first time

https://mp.weixin.qq.com/s/wWQhZIdchK1i_X4pZkjSZQ

19. Li Xiang’s “popular product” strategy: exploring the essence of user needs, not user-centered

https://mp.weixin.qq.com/s/UMEBONbjyGKGufk3F2R04g

20. The ideal X01 configuration with luxurious hardware configuration is suspected to be exposed

https://chejiahao.autohome.com.cn/info/8765333

21. Ideal ONE rear-ended a large truck again! The police have issued a notice claiming that Ideal ONE is solely responsible

https://mp.weixin.qq.com/s/vMpB16HF_UyuvTbd_6ViUw

Latest articles about

■Understanding the OSI Model Using Logistics

■Talk about the controversy of the maximum load of 375 kg for new energy vehicles

■What is the car moose test?

■Live Preview | AUTOSAR SOME/IP Technology Interpretation

■AP AUTOSAR Hard-Core Technology (5): Diagnostic Management

■How much does it cost to customize an automotive-grade ECU?

■Live broadcast today | CAN XL International Seminar

■“Customers are not afraid, so what are you afraid of?” - Reflection on the value of static analysis of automotive software

■Detailed explanation of the control algorithm of the electromechanical brake system (EMB) - Taking Tongyu Automobile and Feige Intelligent as examples

■[Opening this week] SAE-AWC 2024 Automotive EEA Innovation Technology Forum | Free registration