AtmosDialogue

01 / The encounter / Pingxiang & Chongqing

Air pollution directly affects everyone’s health.

My fieldwork team and I visit a patient with a nurse at a rehabilitation center — 01Listening to patients and the people who care for them

Listening in Pingxiang and Chongqing

Growing up in Chongqing, a city shaped by heavy industry, I had already experienced the effects of air pollution on my own health.

To understand the environmental challenges beyond my own experience, I visited old mining communities in Pingxiang, Jiangxi, and Chongqing, interviewing 11 families affected by pneumoconiosis.

A fellow volunteer and I listen to an older resident during an interview — 01Hearing how illness changes a family’s life

The hidden cost became personal.

There, I came to understand the hidden cost of industrial development through the families I met. Years of work in heavy dust had left miners with irreversible lung damage. Losing the ability to work meant losing an income; illness pushed families into poverty and disrupted their children’s education.

I use a laptop while speaking with an older resident at a rehabilitation center — 01Turning conversations into practical support

A volunteer leading a group rehabilitation exercise outdoors — 02Rehabilitation care and support beyond a visit

A volunteer and a resident carrying boxed oxygen equipment to a home — 03Bringing equipment to a family’s doorstep

Volunteers talking with a resident beside an oxygen concentrator during a home visit — 04Home visits connect equipment with ongoing support

An oxygen concentrator with a Love Save Pneumoconiosis label — 05Oxygen equipment provided through the support program

Sustaining care beyond a visit.

Listening to families deepened my commitment to practical care. I organized campus fundraising to help provide oxygen concentrators and living subsidies to 16 low-income families.

We continued peer-support and rehabilitation-care activities at four rehabilitation centers. We aimed to provide practical, sustained support, grounded in what families told us they needed.

11 families: Visited and interviewed
16 households: Oxygen equipment and living subsidies
4 centers: Ongoing rehabilitation care

02 / The response / Research & engineering

Making the invisible measurable.

My fieldwork in Pingxiang and Chongqing exposed the limits of citywide air-quality averages. They could not explain the enormous differences in exposure between people living and working in the same city.

In response, I initiated a project and led a team developing low-cost drone air-monitoring equipment and portable breathing aids.

I began independent research into cross-scale urban microclimates and air pollution, building a monitoring and early-warning framework around vulnerable communities. It connects mobile sensing, prediction, and accessible alerts to make local exposure visible. The research was accepted by GTREE 2026.

I assemble and check the drone in the workshop — 01At the workbench · Assembling the drone platform

I work on project code alongside the printer-monitoring interface — 02A closer look · Working with code and the fabrication interface

Connect the city to the street.

My fieldwork made unequal pollution exposure tangible. Fixed stations describe citywide air quality, but they cannot show every street canyon or the conditions around an individual home.

In my research, I combined mobile UAV sensing with physical simulation. Drone measurements locate pollution hotspots; XGBoost forecasts short-term concentration changes. Language-model interaction translates instructions and measurements into accessible explanations, while a mobile app delivers targeted alerts.

I adjust drone components on a table outdoors — 01Field adjustments · Working directly with the drone hardware

I check the drone outdoors beside a laptop and field equipment — 02Ready for the field · Checking the sensing platform

From the workbench to flight.

Drone-based sensing moves monitoring beyond fixed stations. I worked on the sensing platform, prepared it outdoors, and checked the components directly during field trials.

Reconnaissance, mapped concentration data, forecasting, and language-model interpretation form one workflow for investigating local pollution sources.

Field demonstration · 13 seconds

A simulated source. An actual flight.

Our demonstration begins with a destination-to-coordinates command, followed by a drone taking off beside a simulated pollution source. It shows how our monitoring instructions translate into field actions.

I assemble the sensor prototype — 01Sensor assembly · Building the measurement system

I work with a 3D printer in the laboratory — 02Model fabrication · Preparing the physical experiment

I arrange the buildings of the physical street-canyon model — 03The physical model · Bringing the street canyon to the lab

Build a smaller world. Study what moves.

I used RANS and a k–ε turbulence model to trace vehicle emissions and pollutant dispersion across a 0.15 km² street-canyon environment. With a sensor-equipped, 3D-printed model, I compared stable and unstable atmospheric conditions with the CFD results.

I connected sensor assembly, model fabrication, and physical testing in the experimental platform, using its measurements to compare predicted concentration changes with observations.

0.15 km²: Urban street-canyon model
< 3 min: Warning response time
0.34 m: Simulated source localization

These are the performance results I report in my paper, alongside the full experimental conditions.

02.6 / Paper & project record

The work behind
the numbers.

In my paper, I document the methods, experimental conditions, and results of my research. The paper was accepted by GTREE 2026.

02.1 / PAPER · PDFRead the full research paper 02.2 / PRESENTATION · PPTXRead the project presentation 02.3 / RESEARCH POSTERExplore the complete framework

I present my air-quality monitoring research beside its poster and experimental model — Presenting the research · The framework, experimental model, and results

Research poster / The complete framework

Click to enlarge

Research poster: Large Language Model Empowered Cross-Scale Framework for Urban Air Pollution Monitoring, Diffusion Analysis and Early Warning — Research framework · Methods, experiments, and reported results

Presentation / 10 slides

Selected slides · Click to enlarge

Presentation slide 1: AtmosDialogue. Research overview · LLM-enhanced CFD and drone sensing across urban scales. — 01 / AtmosDialogue

Presentation slide 4: Innovation framework: macro to micro. A shared framework for UAV monitoring and physics-informed interpretation. — 04 / Innovation framework: macro to micro

Presentation slide 6: Macro-scale environment results. Drone reconnaissance, pollution heatmaps, and XGBoost forecasting. — 06 / Macro-scale environment results

Presentation slide 9: Micro-scale environment results. Concentration changes, experimental verification, and the prediction workflow. — 09 / Micro-scale environment results

View all 10 slidesShow fewer slides

Presentation slide 2: Research background and issues. Personal experience, weather-station coverage, and the need for local monitoring. — 02 / Research background and issues

Presentation slide 3: Existing works and limitations. The gap between regional coverage and street-level exposure. — 03 / Existing works and limitations

Presentation slide 5: Macro-scale environment monitoring. UAV hardware, data transmission, and language-model interaction. — 05 / Macro-scale environment monitoring

Presentation slide 7: Micro-scale environment monitoring. Vehicle-emission modeling, street-canyon geometry, and CFD simulations. — 07 / Micro-scale environment monitoring

Presentation slide 8: Experimental platform and 3D-printed microenvironment. A physical model and ESP32-S3 sensor module for atmospheric experiments. — 08 / Experimental platform and 3D-printed microenvironment

Presentation slide 10: Conclusion and future applications. Potential applications in city planning, emergency response, and health guidance. — 10 / Conclusion and future applications

03 / The realization / Beyond technology

A warning cannot undo a lifetime.

Our fieldwork team discusses our investigation in an office — 01Staying with the questions a device cannot answer

Seeing the limits of a technical response

“Even the most precise technology can address future risks. It cannot heal suffering that has already happened.”

The longer I worked with miners, the clearer the limits of my technical response became.

Equipment could ease a difficult breath. A monitoring system could warn of a pollution risk. Neither could, on its own, change the institutional barriers these families faced.

In our conversations, miners described missing employment records and difficulty finding a formal application pathway. Even with a diagnosis, they struggled to secure work-injury compensation and long-term medical support.

Volunteers listening and taking notes during a fieldwork conversation — 01Understanding the barriers families face

My fellow volunteers and I record information during a community fieldwork interview — 02Documenting families’ experiences and needs

From technical innovation to rights advocacy.

What they needed extended beyond monitoring tools: recognition of their work, access to their rights, and a dependable system of support. I realized that addressing environmental injustice meant engaging with the exclusion that kept them unprotected.

I documented the difficulties miners faced, contacted local administrative departments, and helped families organize application materials and navigate policy channels in pursuit of medical subsidies and support for their rights.

04 / The voice / United Nations climate conference

From a mining town to a global conversation.

I speak at the UN climate conference in Bonn — 01Bringing the mining communities’ experiences to Bonn

Bringing the mining communities’ experiences to Bonn

Helping one household mattered. But individual support and case-by-case advocacy could not, alone, change the wider invisibility of these communities.

I sought opportunities to bring their experiences into public discussion, eventually attending the United Nations climate conference as the only high-school youth representative from China.

A wide view of the United Nations climate conference hall in Bonn — 01Inside the conference hall · UNFCCC SB62, Bonn

Speaking for the right to clean air

At UNFCCC SB62 in Bonn, I placed mining families’ breathing difficulties and struggles for recognition at the center of my speech. I called for clean air and health to be recognized as basic human rights, and for global climate governance to hear people too often left outside its conversations.

I exchange ideas with participants at the climate conference — 01Dialogue beyond the speech

I speak with another participant inside the Bonn conference venue — 02Continuing the exchange at the conference

I take part in a climate conference session — 03Listening across perspectives

A conversation that continues beyond the speech

I spoke to carry their concerns further: toward public recognition, stronger protection, and lasting change for communities living with environmental injustice.

AVD certificate recognizing my participation at UNFCCC SB62 in Bonn from 23 to 26 June 2025 — 01Certificate of participation · AVD delegation

A record of participation

I participated in UNFCCC SB62 in Bonn with the AVD delegation from 23 to 26 June 2025.

The certificate records my participation in the climate negotiations and the delegation’s activities.