Against the backdrop of the profound reconstruction of the travel ecosystem by global intelligent driving and new energy vehicle technologies, the higher education system is confronted with a structural contradiction of "accelerated technological iteration" and "lagging talent cultivation". According to data from the Society of Automotive Engineers of China, the demand for compound talents in the field of intelligent driving is continuously growing at an average annual rate of 35%. However, at present, universities still face three core challenges in the discipline construction of intelligent connected vehicles:

1. The absence of the industry-education collaboration mechanism：

The teaching content is disconnected from the actual demands of the industry. The course coverage in cutting-edge technology fields such as advanced autonomous driving algorithm design and multi-sensor data fusion is insufficient, making it difficult to meet the industry's demand for cultivating innovative talents.

‌2. Weak engineering verification system:

Real vehicle tests under extreme working conditions not only pose high safety risks but also have high construction costs, making it difficult for colleges and universities to build a complete engineering verification system and students lack a systematic practical training environment.

‌3.Insufficient interdisciplinary integration:

There are disciplinary barriers in the teaching process of related majors such as mechanical engineering, electronic information, and computer science. There is a lack of effective interdisciplinary collaboration mechanisms and engineering practice platforms, which restricts the realization of the goal of cultivating compound talents.

Relying on its profound accumulation of industrial-level engineering practice, VCARSYSTEM innovates and develops universities AD HIL(Autonomous Driving Hardware In the Loop,Autonomous driving hardware-in-the-loop virtual test systemTeaching platform). This platform, with its closed-loop solution of "empowering education with technology and driving teaching with scenarios", systematically addresses the challenges in the field of intelligent connected vehicle education in colleges and universities, achieving efficient transformation of advanced industrial technologies into educational scenarios, and providing strong support for colleges and universities to build an innovative teaching system that deeply integrates theory and practice.

The VCARSYSTEM AD HIL teaching platform

The VCARSYSTEM AD HIL teaching platform has constructed an engineering practice environment that integrates virtual and real elements. Through the multimodal sensor simulation module and the high-level autonomous driving algorithm framework, it fully presents the entire technical chain of vehicle perception, decision-making, and execution, achieving a closed-loop simulation from environmental perception to dynamic control.

The VCARSYSTEM AD HIL teaching platform deeply integrates the hardware and software ecosystems. With the VCARSYSTEM MINI HIL desktop-level HIL (Hardware-in-the-loop) platform as the core carrier, it integrates the intelligent driving injection unit and the autonomous driving domain controller, and is equipped with the Carla scene simulation software to construct a virtual simulation environment. All components work together to build a technical architecture that combines the virtual and the real. Through the closed-loop link of data interaction and algorithm verification, the full-process simulation from perception and decision-making to control and execution is achieved. The details of its system architecture are as follows:

Data closed-loop process:

• Visual signal stream

Carla scene software renders HDMI data → the intelligent driving injection unit converts it into raw camera data (GMSL video stream) → the autonomous driving domain controller

• Sensor data stream

Carla outputs combined inertial navigation/body attitude/millimeter-wave radar data →MINI HIL platform receives the signal → converts it into CAN message through the bus interface card VCI8 → Autonomous driving domain controller

• Lidar data stream

Carla outputs lidar point cloud data → sends UDP data packets over Ethernet → Autonomous driving domain controller

• Control feedback flow

The autonomous driving domain controller sends control instructions (CAN messages) →MINI HIL parsing → dynamic model calculation → transmits back to the Carla scene software through the self-developed interface model to form a complete closed loop.

The composition of the AD HIL

Dynamic model:

Open multi-dimensional interfaces, adjust vehicle dynamics parameters with one click, and respond quickly to the development needs of different vehicle models

The VCARSYSTEM AD HIL teaching platform replaces the Carla native dynamics model through the self-developed 14-degree-of-freedom (14DOF) vehicle dynamics model based on Simulink.

After engineering verification, this technical solution can significantly improve the testing efficiency of autonomous driving algorithms in complex scenarios, effectively solve key problems such as lagging data transmission and insufficient model accuracy in traditional simulation systems, and provide a high-fidelity and low-latency simulation testing environment for teaching and research in colleges and universities.

The 14-degree-of-freedom vehicle model adopted by the VCARSYSTEM AD HIL teaching platform is highly comprehensive and open, comprehensively covering key modules such as the interface model, driver model, steering system, braking system, powertrain and chassis. This model supports kinematic simulation of vehicles in three dimensions: longitudinal, transverse and vertical.

Grant users extensive and flexible autonomous setting space.

Users can make personalized adjustments to the model's various parameters, module characteristics, and simulation conditions based on different teaching scenarios, research needs, or testing objectives,

realize customized simulation experiences,

providing powerful and flexible technical support for teaching and research in the field of intelligent driving vehicles.

VCARSYSTEM opens all interfaces of the dynamic model and is equipped with a convenient interface modification function. Users can easily complete parameter adjustment and model configuration optimization in the visual operation interface.

For the demand of model replacement, users can directly import the Simulink model with the help of the Kunyi VCAR EM engineering configuration software platform.Achieve efficient and seamless switching of the dynamic model.

Simulation software:

Universities can conduct secondary development based on the open-source code of Carla to build a high-precision scene library and reduce the cost of research and development verification

Relying on the Carla open-source simulation engine, a high-precision digital twin scene library is created to deeply reproduce complex driving conditions such as rain and fog, and low light at night. Users can efficiently verify the core algorithms of autonomous driving and intelligent driving assistance functions such as adaptive cruise control (ACC) and automatic emergency braking (AEB) by simulating diverse real scenarios and extreme boundary conditions. This platform, with a secure and controllable virtual environment, realizes full-domain coverage testing of extreme working conditions, complex scenarios and system failure modes, effectively avoiding the risks of real vehicle testing and significantly reducing the cost of R&D verification.

Autonomous driving domain controller:

The full-stack open-source architecture enables independent secondary development. The modular design lowers the threshold for research and development and enhances the capabilities of programming development and simulation testing

The VCARSYSTEM AD HIL teaching platform has elaborately constructed an open development framework with the ROS2 distributed architecture and integrated an autonomous driving domain controller based on the Autoware open-source algorithm. 

This platform provides users with a highly flexible operation space. Users can optimize and test specific modules, such as the planning control module and the recognition algorithm module, in a targeted manner according to their own research directions. Thanks to the framework built on open-source algorithms, all modules within the platform support users to conduct independent secondary development. 

This open design not only enables students to deeply understand the core principles of the autonomous driving system, but also effectively exercises their programming and development skills as well as simulation and testing capabilities in practice, laying a solid foundation for cultivating professional talents that meet the development needs of the intelligent connected vehicle industry.

VCARSYSTEM Automated Testing Software VCAR EA:

Based on the dynamic scene generalization and high-fidelity simulation capabilities of Python scripts, it realizes full-chain automated testing and efficiently generates massive multimodal data.

VCAR EA supports users in writing test cases to cover diverse scenarios and working conditions, and realizes the dynamic adjustment of scene parameters and scene generalization through Python scripts.

VCARSYSTEM AD HIL Teaching Platform24-hour uninterrupted testing is achieved by using automated scripts, and parallel testing is realized by running different test cases on multiple servers. In automated testing, different conditions such as weather and lighting are set accordingly by inputting different parameters.

Easily achieve scene generalization and efficiently generate massive multimodal data, providing strong support for the rapid iteration of algorithms and the precise optimization of parameters.

This platform deeply integrates high-fidelity dynamic models with multi-physics field coupling simulation technology, and can highly accurately reproduce complex scenarios such as vehicle aging effects, harsh environmental interference, and abnormal sensor states. Through the simulation of these complex scenarios, the platform can comprehensively verify the full-chain collaborative performance of the perception, planning and control modules.

Ensure that the intelligent driving system has highly reliable functional performance under various complex conditions, providing a solid guarantee for the teaching, research and application of intelligent driving technology.

VCARSYSTEM Desktop HIL Platform MINI HIL:

High real-time system, achieving closed-loop verification of high-level autonomous driving algorithms

The VCARSYSTEM desktop HIL platform MINI HIL simulates the vehicle's motion characteristics in real time through a 14-degree-of-freedom vehicle dynamics model and supports closed-loop verification of high-level autonomous driving algorithms.

The bus interface card VCI8 equipped on the device has a powerful communication capability, integrating 8 CAN, 8 LIN and 1 FlexRay interface, which can achieve seamless connection with the real controller. Through high-precision simulation and analysis of data from multiple types of sensors such as combined inertial navigation and millimeter-wave radar,

help students systematically master core industry skills such as multi-sensor data fusion technology and vehicle control algorithm development, and quickly transform theoretical knowledge into engineering practice ability.

Camera video injection unit:

Convert the Carla virtual scene into GMSL video streams in real time to achieve high-precision video data injection

The video injection unit of VCARSYSTEM Camera can convert the HDMI virtual scenes rendered by Carla (covering multiple camera perspectives) into camera data streams (GMSL video streams) in real time and precisely inject them into the autonomous driving domain controller.

Through immersive practice, students are helped to deeply master core development skills in the industry such as camera calibration and perception algorithm optimization, and quickly achieve the leap from theoretical learning to engineering application.

Application of AD HIL Teaching

The VCARSYSTEM AD HIL teaching platform is deeply in line with the demands of diverse teaching scenarios. By constructing a systematic and full-process experimental teaching system, it provides a one-stop solution for the cultivation of professional talents in the field of autonomous driving. The platform focuses on the following core teaching scenarios, comprehensively empowering the integration of theory and practice as well as the improvement of technical capabilities:

1. Teaching and verification of autonomous driving algorithms

Construct the core testing capability system for the teaching and verification of autonomous driving algorithms in colleges and universities:

Simulation scenario construction capability: 

Build virtual simulations covering complex traffic scenarios such as severe weather, traffic congestion, and sudden obstacles, and help students master algorithm robustness analysis and optimization techniques through extreme condition tests.

Multi-sensor fusion capability: 

Integrate data from multiple sensor sources such as cameras, millimeter-wave radars, and lidars to verify the collaborative effects of perception, decision-making, and control algorithms, and deepen students' understanding and cognition of the collaborative mechanisms of algorithms.

Hardware-in-the-loop simulation capability :

Provide real in-vehicle hardware interfaces (such as ECU controllers) to achieve seamless connection of algorithms from virtual simulation to real vehicle applications, ensuring the deep integration of teaching practice and engineering applications.

2. Scientific research innovation and project incubation

Interdisciplinary collaborative innovation: 

Integrating resources from multiple fields such as vehicle engineering, artificial intelligence, and communication technology to support students in conducting scientific research and technological exploration on cutting-edge topics in autonomous driving.

Deep integration of industry, academia and research:

Build a bridge for cooperation between schools and enterprises, support universities in providing forward-looking technological research services to enterprises, and accelerate the efficient transformation of scientific research achievements from laboratories to industrial applications through joint research and development, technology transfer and other means, promoting the mutual empowerment of technological innovation and industrial upgrading.

3. Competition training and laboratory construction

Competition ecosystem cultivation:

Build a professional hardware platform and provide in-depth empowerment World Vocational Skills Competition We provide comprehensive support for university teams participating in authoritative competitions. With real vehicle dynamics models and real-time debugging environments, it has significantly enhanced students' engineering practice abilities and has helped many colleges and universities achieve excellent results in various vocational skills competitions.

Innovation of Practical Training Courses:

We have meticulously crafted modular practical training courses such as autonomous driving system testing and data closed-loop training, and established a three-in-one teaching system of "theoretical foundation - simulation simulation - practical operation", precisely matching the demands of cutting-edge talents in the industry and achieving seamless integration of teaching content with industrial practice.

Laboratory Upgrade Project:

We will fully assist universities in building intelligent connected vehicle laboratories and high-level autonomous driving research and development centers, improve the infrastructure for collaborative innovation among industry, academia and research, and provide solid hardware support and innovation platforms for technological research and development, talent cultivation and the transformation of achievements.

The VCARSYSTEM AD HIL teaching platform takes teaching, scientific research and industry as the three major engines, and builds a closed-loop ecosystem of "cultivating innovative thinking through basic teaching, forging core capabilities through technological breakthroughs, and accelerating the transformation of achievements through industrial implementation".

By deeply integrating theory with practice, academia with industry, we aim to lay a solid foundation for cultivating outstanding compound talents in the field of intelligent driving who possess both systematic architecture thinking and proficiency in engineering applications, and to promote the synchronous development of the discipline and industry innovation.

Up to now, the VCARSYSTEM AD HIL teaching platform has been deployed and applied in many universities across the country and has received wide recognition. Looking ahead, VCARSYSTEM will deepen the school-enterprise collaborative education mechanism with a more open attitude, focus on the deep integration and precise matching of the education chain, talent chain and industrial chain, and continuously improve the integrated innovation system of "industry-university-research-application". By constantly optimizing teaching resources and technical services, we cultivate and supply high-quality core talents with both cutting-edge innovative thinking and outstanding practical abilities to the intelligent automotive industry, fully promoting the high-quality development of the industry and technological innovation.