Industry Collaboration
For manufacturers and equipment makers — deploying industrial AI algorithms onto real production lines, from joint R&D to technology transfer
Collaboration Models
Four Ways to Partner With Us
Whichever stage your need is at — exploration, technical breakthrough, or industrialization — there is a fitting path
- 01
Joint R&D
Forming joint teams to research and iterate algorithms around a company's specific real-world production-line challenges, with shared outcomes
- 02
Contract Development
Companies define clear technical requirements; the lab independently develops the algorithm/system and delivers a deployable solution against milestones
- 03
Technology Transfer & Licensing
Licensing existing patents and algorithms, with follow-on development support, to accelerate industrialization
- 04
Joint Lab Partnership
Co-building a joint laboratory or technology station with on-site, long-term collaboration that continuously delivers production-line capabilities
Case Study · Rail Transit Operations
Rail Transit Intelligent Diagnosis Agent
Multimodal Industrial Spatiotemporal Large Model for Rail-Transit Intelligent Diagnosis
Built for high-speed and urban-rail operators, turning images, time series, and text logbooks into one interpretable diagnostic chain via a large-model agent. View Project Details →Real Pain Point
Rail-corrugation inspection lacks a unified standard and is prone to missed/false detections; data spans images, time-series waveforms, and text logbooks, which traditional methods struggle to fuse
Technical Approach
A large-model agent diagnostic framework combining visual autocorrelation feature extraction with time-series waveforms and text-logbook reasoning for automated, fine-grained diagnosis
Outcome
Achieved 96% visual rail-corrugation recognition accuracy, while automatically identifying and distinguishing complex operating conditions such as turning and impact, with full interpretability
Case Study · Domestic Edge-Intelligent Control
Fully Domestic Edge-Intelligent Control System
Domestic PLC+GPU Edge-Intelligent Control Hub
For equipment makers seeking domestic-technology substitution and real-time closed-loop control, bringing the "brain" onto the production line itself with panoramic visualized equipment management. View Project Details →Real Pain Point
Cloud-based diagnostic chains have high latency — the gap between "detecting an anomaly" and "stopping the line" is often second-scale — and core software/hardware remain heavily dependent on foreign suppliers
Technical Approach
A fully domestic PLC+GPU dual-core edge system that compresses sensing, health assessment, and control response into a single local closed loop
Outcome
Closed-loop latency cut from second-scale to under 50 milliseconds, fully domestic and self-sufficient across the chain, backed by hardwired safety interlocks
Collaboration Process
How to Get Started
From first contact to project delivery — a clear, controllable process that lowers the cost of a company's decision
Step 1: Initial Discussion
Share your requirements and production-line scenario by email; both sides assess feasibility and fit
Step 2: Sign an NDA
Once core technical needs and data boundaries are clear, sign a non-disclosure agreement to protect your processes and data
Step 3: Proposal & Project Setup
The joint team defines the technical plan, milestones, and budget, then formally launches the project
Complete: Delivery & Acceptance
Deliverables are shipped against milestones with on-site integration; after acceptance, the collaboration can continue iterating or move to a long-term partnership
Start an Industry-Academia-Research Partnership — With One Email
Email: fengjsh7[at]mail[dot]sysu[dot]edu[dot]cn | Address: Room 518, East Building, Science Park, Sun Yat-sen University, 66 Gongchang Road, Guangming District, Shenzhen, Guangdong Province