Intelligent Connected Vehicle (ICV)

An Intelligent Connected Vehicle (ICV) refers to a new generation of vehicles equipped with advanced on-board sensors, controllers, and actuators, seamlessly integrated with modern telecommunications and networking technologies. These vehicles facilitate intelligent data exchange and sharing among drivers, vehicles, road infrastructure, and cloud systems, offering complex environmental perception, intelligent decision-making, and collaborative vehicle control, ultimately designed to take over manual driving operations.

Conceptually, ICVs encompass two core dimensions: "intelligence" and "connectivity". This requires the vehicle to possess autonomous environmental perception and automated driving control capabilities, while enabling real-time data communication with external systems. If electrification represents the "first half" of the automotive industry's transformation, intelligent connectivity is the decisive "second half" that will determine the ultimate winners.

Technical Architecture and "Vehicle-Road-Cloud-Network" Synergy

The technical framework of an ICV is best understood as a"Vehicle-Road-Cloud-Network" integrated collaborative architecture. On-board sensors (cameras, millimetre-wave radars, LiDAR, etc.) serve as the vehicle's "eyes", scanning the surrounding environment; high-performance computing units act as the "brain", processing vast datasets to execute driving decisions; drive-by-wire chassis systems (brake-by-wire, steer-by-wire, etc.) function as the "hands and feet", translating digital commands into physical vehicle manoeuvres. Meanwhile, 5G and cellular communication networks act as the "central nervous system", supporting seamless real-time data exchange between vehicles, transport infrastructure, and cloud platforms.

In terms of technological roadmaps, the coordinated development of stand-alone vehicle intelligence and cloud-network connectivity has been established as China's core strategy. The industry is shifting from traditional modular architectures towardsend-to-endsolutions, consolidating perception, path planning, and vehicle control into a unified neural network model to map raw sensor data directly to control commands. Concurrently, automotive software architectures are evolving from "Software-Defined Vehicles" (SDV) to "AI-Defined Vehicles" (AIDV), embedding generative AI and foundation models into the core platform to enable proactive perception, autonomous decision-making, and continuous over-the-air evolution.

At a deeper level, ICVs function ascomplex Cyber-Physical Systems (CPS)that deeply integrate automotive engineering, transport networks, computing, and telecommunications. China's blueprint is built upon five foundational platforms: theCloud Control Platform, High-Definition Dynamic Map Platform, Intelligent On-Board Terminal Platform, Computing Platform, and Cybersecurity Platform—together establishing a unified "Human-Vehicle-Road-Cloud" ecosystem.

Current Landscape and Key Challenges

China's ICV sector has entered a new phase of large-scale commercialisation. Passenger cars equipped with Level 2 (L2) combined driver assistance systems now account for over 60% of total new vehicle sales. High-compute SoC chips and smart drive-by-wire chassis components have achieved mass-production adoption, while human-machine interface (HMI) and cooperative perception technologies have reached world-class standards.

In terms of testing and validation, over 35,000 kilometres of public test routes have been gazetted nationwide, supported by more than 11,000 intelligent roadside units (RSUs) and 17 national-level proving grounds, with cumulative test mileage exceeding 120 million kilometres. Across five major metropolitan clusters—including Beijing-Tianjin-Hebei, the Yangtze River Delta, and the Pearl River Delta—various pilot programmes have been rolled out for smart mobility, autonomous vehicle platooning, and long-haul trunk freight transport.

Nevertheless, the rapid proliferation of ICVs introduces critical challenges. Safety remains an unavoidable priority—when confronted with corner cases or rare edge scenarios not fully covered in AI training datasets, autonomous systems may fail due to a lack of baseline knowledge, while adverse weather and sensor noise can cause misdetections and incorrect driving inputs. On the regulatory front, the industry is transitioning from "passive compliance" to "active resilience", with several key technical standards concerning public safety shifting from voluntary recommendations to mandatory national standards.