Sensor Web is a Mozilla project to develop tools to make it easy to deploy sensors to collect real-time data and publish it via the web. Initially, the plan is to collect PM2.5 data, which measures air quality (specifically the amount of <2.5 micrometer particlate pollution in the air).
Questions
What existing work can we leverage in this project?
Who can we team up with?
Existing work
Sensor Web Enablement
"Sensor Web" is actually the name of a NASA project started in 1997 which does something very similar to our project. Since then, it's been standardized as "Sensor Web Enablement" and managed by the Open Geospatial Consortium. OGC has a huge amount of research on this topic already, and has established standards being used by a variety of organizations, including NASA, Google, Oracle, MIT, and many US government departments. They also work with standards bodies including ISO and W3C.
SensorThings
SensorThings is a new JSON/REST-based IoT API (it just passed voting in February 2016). This may prove useful for other teams working on IoT, as it's an existing standard that seems to fit closely with our needs. Much of the SensorThings work is being done by SensorUp, a Canadian company working to provide open IoT standards. As part of the Eclipse IDE, they've created Whiskers, a Javascript-based client that can speak to Raspberry PIs via the SensorThings API. Whiskers seems to be very early in development, however.
Open Sensor Hub
Open Sensor Hub is an existing project targeting OGC standards and designed to run on devices like Raspberry Pi. It also sounds like a close fit for our needs. It's even licensed under the Mozilla Public License! Open Sensor Hub appears to be fairly mature (they released version 1.0 in October 2015), although I haven't deeply reserached its feature set yet.
Measuring PM2.5
There are lots of projects on GitHub for measuring PM2.5, with a very strong emphasis on China. There's not much in other parts of the world; how much relevance does this have globally?
In particular, http://pm25.in is referenced in some GitHub projects as having an API to fetch PM2.5 data. While I can't read Chinese, there appears to be a lot of data there already. This service seems to already be providing something very similar to our initial plan.
Conclusions
I strongly recommend that we use OGC standards when developing our software; it appears to be widely-accepted as the standard in this space (research-oriented IoT). I also think we should take a closer look at either Whiskers or Open Sensor Hub, since they sound like they meet our needs. Open Sensor Hub appears to be much more mature than Whiskers, so I'd recommend looking at that first. It may also be useful to work with these groups directly to consolidate our efforts.
Further, I think we should look at partnering with researchers themselves. My understanding is that Mozilla's Sensor Web is primarily targeting hackers/makers, but a cheap, easy-to-use sensor platform would also be very attractive to research teams (e.g. academics). While I think we should still make this work accessible to people performing "citizen science", professional researchers have the skillset to produce more-reliable data. We shouldn't ignore them.
Finally, the SensorThings API may prove useful to IoT work on other trains.
Sensor Web
Summary
Sensor Web is a Mozilla project to develop tools to make it easy to deploy sensors to collect real-time data and publish it via the web. Initially, the plan is to collect PM2.5 data, which measures air quality (specifically the amount of <2.5 micrometer particlate pollution in the air).
Questions
Existing work
Sensor Web Enablement
"Sensor Web" is actually the name of a NASA project started in 1997 which does something very similar to our project. Since then, it's been standardized as "Sensor Web Enablement" and managed by the Open Geospatial Consortium. OGC has a huge amount of research on this topic already, and has established standards being used by a variety of organizations, including NASA, Google, Oracle, MIT, and many US government departments. They also work with standards bodies including ISO and W3C.
SensorThings
SensorThings is a new JSON/REST-based IoT API (it just passed voting in February 2016). This may prove useful for other teams working on IoT, as it's an existing standard that seems to fit closely with our needs. Much of the SensorThings work is being done by SensorUp, a Canadian company working to provide open IoT standards. As part of the Eclipse IDE, they've created Whiskers, a Javascript-based client that can speak to Raspberry PIs via the SensorThings API. Whiskers seems to be very early in development, however.
Open Sensor Hub
Open Sensor Hub is an existing project targeting OGC standards and designed to run on devices like Raspberry Pi. It also sounds like a close fit for our needs. It's even licensed under the Mozilla Public License! Open Sensor Hub appears to be fairly mature (they released version 1.0 in October 2015), although I haven't deeply reserached its feature set yet.
Measuring PM2.5
There are lots of projects on GitHub for measuring PM2.5, with a very strong emphasis on China. There's not much in other parts of the world; how much relevance does this have globally?
In particular, http://pm25.in is referenced in some GitHub projects as having an API to fetch PM2.5 data. While I can't read Chinese, there appears to be a lot of data there already. This service seems to already be providing something very similar to our initial plan.
Conclusions
I strongly recommend that we use OGC standards when developing our software; it appears to be widely-accepted as the standard in this space (research-oriented IoT). I also think we should take a closer look at either Whiskers or Open Sensor Hub, since they sound like they meet our needs. Open Sensor Hub appears to be much more mature than Whiskers, so I'd recommend looking at that first. It may also be useful to work with these groups directly to consolidate our efforts.
Further, I think we should look at partnering with researchers themselves. My understanding is that Mozilla's Sensor Web is primarily targeting hackers/makers, but a cheap, easy-to-use sensor platform would also be very attractive to research teams (e.g. academics). While I think we should still make this work accessible to people performing "citizen science", professional researchers have the skillset to produce more-reliable data. We shouldn't ignore them.
Finally, the SensorThings API may prove useful to IoT work on other trains.