CINEVE | Physical AI Will Reshape the Automotive Production Model | Frontop Digital Creative Technology Ye Mi Xing Shares New Path for Smart Manufacturing

At the AI-powered Automotive Closed-door Seminar held concurrently with the China International New Energy Vehicle Technology, Parts, and Services Exhibition (CINEVE), Ye Mi Xing, Solutions Director of Guangzhou Frontop Digital Creative Technology Co., Ltd., delivered a keynote speech on the empowerment of the automotive industry through physical AI. He systematically shared the digital and intelligent upgrade path and implementation results of Frontop’s industrial physical AI in automotive research and development, as well as production processes, which attracted widespread attention from the supply chain and automotive companies on-site.

1. Three Major Realistic Challenges Facing the High-Quality Development of the Automotive Industry

Ye Mi Xing mentioned that the automotive industry is accelerating towards electrification, intelligence, and connectivity. However, during this process, it still faces prominent issues. "The relatively long research and development cycles, high manufacturing costs, and the difficulty in quality control after the vehicle is launched are common pain points that the industry is focused on."

He stated that under the promotion of the "Artificial Intelligence + Manufacturing" policy, the upgrade of intelligent manufacturing in the automotive industry has become urgent, and physical AI will be the key technological direction to break through traditional models and empower automotive manufacturing. He believes that by 2026, AI technology will transition from "virtual" to "physical."

2. From Generative AI to Physical AI: Transitioning from Content Generation to Physical World Interaction

He clearly pointed out that traditional generative AI mainly focuses on content generation such as text, images, and code, driven by data analysis. "It lacks an understanding of physical details and production processes, and cannot truly interact with production lines, AGVs, or materials." The core of physical AI lies in understanding the underlying logic and process rules of the world. It can detect production line issues through multimodal perception, make decisions, and drive robots and AGVs to perform material scheduling and process adjustments, thus achieving deep interaction with the physical world.

3. Building a Closed Loop with a Four-Layer Architecture to Create a Digital Parallel World for the Automotive Industry

He introduced that Frontop Digital Creative Technology has achieved a full-process closed loop for physical AI through a four-layer product system: The bottom layer is the perception and data foundation, connecting IT/OT multi-source data. The upper layer is the AI 3D digital twin simulation engine, constructing a virtual world that adheres to real physical rules such as gravity, friction, and collision. The next layer is the physical AI simulation decision-making platform, integrating large models and intelligent agents to enable autonomous decision-making. The top layer consists of embodied intelligence and execution devices, completing physical actions in reality.

Through this system, companies can achieve full-link management from perception to decision-making and execution, creating a digital parallel world that is synchronized with the real production line.

4. Industry-Academia-Research Collaboration and Technological Accumulation to Support Scene-Based Implementation 

Ye Mi Xing stated that the company has accumulated a billion-level 3D industrial model asset library, enabling the rapid construction of production lines, warehouses, and equipment scenarios, and achieving environmental simulation and production data integration. At the same time, Frontop Digital Creative Technology has established joint laboratories with the Shenzhen Artificial Intelligence and Robotics Research Institute, relevant laboratories at Peking University, and Guangdong University of Technology. These collaborations focus on technological breakthroughs in embodied intelligence, multimodal perception, industrial simulation, human-machine collaboration, and other areas.

5. Clear Boundaries of Human-Machine Collaboration: AI Handles Repetitive Tasks, Humans Focus on Critical Judgments

In terms of application implementation, he focused on sharing the core scenarios of physical AI in the automotive industry: 

• Physical AI Control Room: Serving as the digital brain, it enables full panoramic monitoring of the production line, AI diagnostics, and predictive maintenance of equipment.

• Production System Simulation: Supporting multi-model mixed-line production, production line switching, and simulation optimization before line expansion or upgrades.

• Physical AI Virtual Testing: It can simulate extreme and rare scenarios such as flooding, fires, and tire blowouts, addressing the limitations of real-vehicle testing.

• Warehouse-Logistics Collaboration: Connecting the entire chain from production line to warehouse to logistics, achieving intelligent scheduling of AGVs/RGVs.

• Connected Traffic Testing and Monitoring: Real-time tracking of test vehicle status, enabling safety event early warnings and assessments.

He emphasized that physical AI has already been implemented and validated in benchmark projects such as public transportation groups. It is not only a technological upgrade but will also drive a fundamental transformation in automotive production models. "We firmly believe that physical AI will deeply empower intelligent automotive manufacturing, driving high-quality development in the industry through technological innovation."

Relevant companies visited the Frontop Digital Creative Technology booth at the Yasn Exhibition.

Exciting Event Preview 

The AI-powered Automotive Closed-door Seminar will precisely match the real needs of automotive companies, unite industry consensus, and promote in-depth integration of cutting-edge technologies such as physical AI, digital twins, and the fusion of virtual and physical worlds with automotive R&D, manufacturing, and testing scenarios. This will lay a solid foundation for the large-scale implementation of AI technologies through collaborative partnerships. 

In response to the exclusive needs of Changan Automobile, we will visit Changan in May to host the AI-powered Automotive Industry Deep-Dive Matchmaking Conference and Live Technology Demonstration, focusing on core areas such as physical AI, intelligent manufacturing, and virtual-physical testing. We welcome companies in the AI + automotive sector with technological highlights and solutions to participate, fostering technology transformation and exploring new opportunities for industry collaboration!