Optical mechanisms
Understand how signals form through sensing and demodulation, and how system characteristics affect observations.
PHYSICS-INFORMED AI + FIBER SENSING
Bring physics into AI. Understand real-world signals.
Acquire field information through distributed fiber sensing and integrate optical mechanisms, propagation principles and engineering constraints into model development, exploring the path from signal analysis to event understanding.
01 / FROM FIBER TO SENSING
One optical fiber. A continuous sensing pathway.
Fiber deployed along infrastructure can do more than carry information: it can reveal changes in the physical world. Distributed fiber sensing captures signals along the route; algorithms and AI models turn those disturbances into event insights that can be located and investigated.
02 / HOW IT WORKS
Send a pulse of light. Read the changes along its path.
In distributed acoustic sensing (DAS), an interrogator launches optical pulses into a single-mode fiber and reads the weak backscattered light. Vibrations or strain affect the optical signal. Analysing these changes and their return timing reveals the location and temporal features of disturbances along the route.
- 01
Launch a pulse
An interrogator connects to the fiber and launches probe light.
- 02
Read the response
Read backscattered light to resolve changes along the route.
- 03
Interpret the signal
Signal processing and model analysis provide insights for field investigation.
Sensing coverage and signal quality depend on the sensing method, installation, coupling and environmental conditions.
03 / PHYSICS × DATA-DRIVEN LEARNING
Help AI interpret signals. Ground it in how they arise.
Similar events can produce different signals under different installation and environmental conditions. OX-SENSE is combining physical knowledge and data-driven learning, incorporating optical sensing processes, propagation characteristics and engineering conditions into model development for a better-grounded understanding of events in complex settings.
PHYSICAL KNOWLEDGEFIELD DATA
EVENT UNDERSTANDING
Propagation principles
Study the temporal, spatial and frequency characteristics of disturbances to provide physical grounding for event representation and model analysis.
Engineering constraints
Consider installation, site coupling, background noise and operating conditions to connect research with practical application challenges.
04 / TECHNOLOGY HIGHLIGHTS
Continuous sensing. Grounded understanding.
Connect the sensing capabilities of fiber with algorithms, models and engineering experience.
Single-mode fiber
Evaluate existing fiber for sensing against available cores, installation and coupling conditions.
Distributed sensing
Capture spatially distributed observations to understand where and how changes occur.
AI event recognition
Combine signal features and application context to study event classification and background interference.
Physics-informed models
Incorporate optical mechanisms, propagation principles and engineering constraints into model development and validation.
05 / BUILT AROUND FIELD CHALLENGES
Advance models and algorithms around field challenges.
Each method must be assessed against specific datasets, test conditions and field validation.
Noise interference
Study how to distinguish background disturbances from target event features, improving the discernibility of useful information.
Limited samples
Explore how physical knowledge and data-driven learning can work together when event samples are limited.
Changing conditions
Study model performance as field conditions change, using cross-scenario validation to assess applicability.
06 / MONITORING AND COLLABORATION
See the event. Understand the next step.
From signals and event locations to investigation records, a monitoring platform gives analysis its field context. OX-SENSE is developing data, models and the platform together around event location, signal review and human investigation. Integration with existing operations systems is assessed against project needs and interface requirements.
- 01
Data
Signals along the route
- 02
Analysis
Algorithm and model outputs
- 03
Events
Locations and event insights
- 04
Investigation
Human judgement and coordinated response
07 / RESEARCH FOUNDATIONS
Explore the research behind the technology.
The following public research involves our founder and collaborating teams. Product capabilities depend on the specific solution and its validation results.
AI-driven distributed acoustic sensing
Review · PhotoniX · 2025 A framework for sensing and intelligent analysis, covering work by multiple teams in the field.
Read the research ↗Learning methods for DAS signal reconstruction
Conference paper · OFC 2025 Explores neural networks for spatial-resolution research. Link to the institutional record and abstract.
Read the research ↗Fiber sensing and AI agents
Institutional research report · SUSTech · 2026 Introduces research into integrating sensing systems with AI agents.
Read the research ↗
FEEL THE WORLD
Define the technical challenge from your site.
Discuss your assets, target events and field conditions to explore sensing methods, model adaptation and a validation pathway.
Discuss technical needs