01
FunSim physics AI simulation training platform
AI+BI dual-core driver, flexible access to massive cloud data, low-code/zero-code approach, achieving lightweight product deployment
Hall A · Booth A507
Frontop Digital Creative Technology Co., Ltd. (Guangzhou)
Guangzhou Fantuo Digital Creative Technology Co., Ltd. (stock code: 301313) was established in 2002 and is the world's leading comprehensive provider of physical AI full-stack prod…
Supply Chain & Components
01
COMPANY PROFILE
Company Profile
Guangzhou Fantuo Digital Creative Technology Co., Ltd. (stock code: 301313) was established in 2002 and is the world's leading comprehensive provider of physical AI full-stack products and integrated solutions. Relying on nearly 1 billion AI in 25 years With profound accumulation of 3D digital model assets, we have successfully launched innovative chain-based multi-functional tools such as FTE digital twin simulation rendering engine, FunSim physical AI simulation training platform, full-stack self-developed real machine data collection and annotation training software and hardware, and embodied robot navigation expansion backpack kit that are in line with domestic Xinchuang. The combination of tools provides comprehensive system support for embodied intelligent autonomous tool chain technology research, innovative development of industry application scenarios, integrated engineering construction of embodied training venues or industrial innovation centers, etc., and creates a series of scenario innovation benchmarks in industries such as intelligent manufacturing, water conservancy and water affairs, energy and power, culture and museums.
02
MAIN PRODUCTS
Key Products
05 ITEMS
02
High quality data set
Provides labeled high-quality data, which can be used out of the box to directly empower embodied intelligent model training. Relying on massive assets such as 1 billion sets of 3D model and scene data, and 10 million+ sets of simulation interaction data, it covers the entire link of data products, customized services and self-developed annotation platforms, shortening the robot training cycle to 1/3 of the traditional one.
03
Embodied robot software and hardware integration services
1. Modular reconstruction: Flexible combinations of robot dogs with robotic arms, AGVs with single arms, etc., standard interfaces can be quickly reorganized, and the threshold is low to adapt to inspection, transportation, and collaboration scenarios. 2. Simulation training: The self-developed physical simulation engine builds a high-fidelity virtual field, and the data is automatically annotated, shortening the real machine debugging time and reducing collection costs. 3. Algorithm delivery: The scenario-based end-to-end pre-training model is plug-and-play and integrates motion planning, perception, and decision-making functions. 4. Full-stack service: one-stop implementation of hardware debugging, integrated deployment, and operation training, and continuous technical optimization and operation and maintenance to ensure stable production.
04
Integrated training ground solution
Currently, the construction of embodied intelligent training ground is faced with four major bottlenecks: "the venue lacks a closed loop first, the scene driver is difficult to copy, the tool gives priority to no data, and multi-modal deep cultivation is difficult to implement." Fantuo Digital Innovation breaks the industry dilemma and combines the four roads into an "embodied intelligent highway" - based on "AI 3D spatial intelligence", it builds a robot digital twin training ground that is "growable, trainable, and easy to implement" and provides full-stack services from high-precision data collection to city-level simulation training
05
Embodied action choreography cloud creation platform
1. Video action imitation: Upload real-life videos, AI extracts skeletal movements and converts them into robot instructions to adapt to dance and teaching demonstrations. 2. Intelligent voice-driven arrangement: text/voice input generates emotional speech, semantically matching action expressions, suitable for speeches and hosting. 3. Visual timeline arrangement: Drag and drop to edit actions and voice materials, and freely connect them to complete story-based performance arrangement. 4. Material library: Built-in standardized actions and voice templates, and supports uploading custom materials to build an exclusive resource library.