Edge AI and low-cost sensors for community disaster preparedness — a resident-driven data system in Hawaii (S&CC) (NSF AI grant $1.5M)

The NSF awarded about $1,497,357 to a project using edge AI, biofabrication, and data science for disaster preparedness and community wellbeing (NSF Award 2531574; program: Smart & Connected Communities / IT Research; Hawaii).

1Infrastructure and health in remote communities

Per the abstract, communities in Hawaii and similar remote, rural, and austere regions often face dynamic local conditions affecting public infrastructure and wellbeing. Events such as sudden changes in water quality or air particulates after a disaster pose significant challenges, and effective monitoring requires timely, localized data that is hard to obtain with existing infrastructure.

2A Smart and Connected Communities design

This Smart and Connected Communities (SCC) project addresses this gap by enabling neighborhoods to become active participants in data collection and analysis of their surroundings.

It collaborates with participants across several Hawaii sites to co-design and build novel, low-cost systems for localized data acquisition, featuring customizable sensors fabricated via advanced manufacturing (3D printing) connected to a powerful, portable data-analysis platform (edge AI).

By making these tools available to local communities, the project enables rapid, localized responses to unforeseen events and provides valuable data for disaster planning and response. Technically, it aims to develop and integrate three core innovations including a low-cost, open-source electronics printer and a scalable platform.

3The order in which residents become the actors

The project does not have researchers measure and inform residents; it sets out to make residents themselves the ones who gather the data.

1. 1Co-designDecide with residents across several sites in Hawaii what is needed
2. 2BuildProduce customisable low-cost sensors by 3D printing and related methods
3. 3Solve on the spotProcess locally on a portable edge-AI platform rather than sending data away
4. 4UseRespond neighbourhood by neighbourhood to sudden changes in water or air after a disaster

In remote and rural places the difficulty is getting data for the right place at the right time. Processing on the spot rather than sending it to a centre is a design made for the period after a disaster, when communication may be cut.