Wera

Smart, personal, realtime urban heat warning system.

Project during Studies

Year: 2024

Duration: 4 Weeks

Team: Ege Seçgin, Stepan Vedunov, Elia Salerno

Role: Conception, Research, User Testing, Prompt Engineering, Electrical Engineering, Web- and Embedded Development


GitHub Repository (Monorepo)

Health Design

Service Design

Interaction Design

Embedded Systems

IoT

Wera project video

Our defined focus

How might we create an accessible, location-based climate risk communication system that provides real-time climate warnings with natural language, informing seniors in cities like Zürich preventively of immediate risks such as high urban temperatures or icy pathways?

High density urban climate sensor network (Meteoblue API)

High density urban climate sensor network (Meteoblue API)

Heat map of Zurich (Meteoblue API)

Heat map of Zurich (Meteoblue API)

As there is a high density urban climate sensor network in Zürich, we contacted the initiators of the project at ZHAW and found that Meteoblue is currently maintaining it, but there is currently not much utilization of it.

Project idea scoping

Project idea scoping

Wera prototype

Wera prototype

Wera prototype internals

Wera prototype internals

Wera interfacing devices

Wera interfacing devices

As the Wera service is device agnostic, we thought of different interfacing methods, of which we built working prototypes for both the Kub and the Fong.

Sketches for the industrial design of the Kub based on radio formfactor inspiration

Sketches for the industrial design of the Kub based on radio formfactor inspiration

CAD model of the Kub

CAD model of the Kub

Industrial design of the Kub

Industrial design of the Kub

Final Prototype

Final Prototype

User testing at an elderly home

User testing at an elderly home

We tested the Kub with two seniors, both of whom easily understood its simple interface—a single button that rotates to adjust volume and clicks for a climate report. The voice output, available in German and English, was also well received. One participant suggested longer reports, while the other recommended adding clothing advice for better weather preparedness.

Findings

Our research highlights a lack of tailored solutions for heat-related risks among the elderly in Switzerland. Initial challenges included interpreting data from the urban climate sensor network and identifying how to deliver warnings effectively, but we were able to solve that withing sprints. Field studies showed that seniors trust personal warnings from family and friends. Using Llama 3 and text-to-speech AI, we personalized location-based alerts, which users perceived as direct and reliable. Removing smartphone dependency further improved accessibility. The project demonstrated strong potential, with future expansions planned for icy pathway and pollen warnings.

ZVV

Localista