Disaster management using edge-of-field computing and low-cost sensors

If there is something that has become much clearer in recent years with climate change, it’s that the magnitude and frequency of natural disasters has increased dramatically in the last 5 years. The Office of Management and Budget in the US Government conducted a preliminary analysis this year and concluded that six major types of natural disasters (coastal disasters, flooding, crop failures, climate-related health disasters, wildfires and building failures)  “were likely to result in annual expenditures of approximately $134 billion and could result in as much as $2 trillion in lost revenue by the end of the century”. 


As a result, there’s been a lot of interest in technologies that can help responses in the aftermath of disasters as well as those that can help detect these disasters early and hence mitigate the damage from them. 


Recently, scientists from OakRidge National Laboratory helped develop a prototype system to detect damaged utility poles from hurricanes and other disasters. This system uses a combination of hardware in the form of drones and sensors together with machine learning algorithms - and can be deployed quickly and efficiently in locations where local infrastructure has been damaged or destroyed. The interesting aspect of this system is that it uses an efficient “edge of field” computing technology that will allow the system to function independently - a real boon in conditions when internet access is unreliable. 


Edge of field computing on on-device computing is typically when machine learning algorithms are deployed on sensors and/or devices and do not connect directly to the internet. The advantage of these systems is that they are small, easily deployed and more robust in conditions with limited resources. They are often more secure as well. The disadvantage is that because of computational limitations since they don’t use the typical cloud-computing resources, algorithms need to be developed that can function with limited data. 


In order to get past that limitation, the team at OakRidge started training the algorithms on the cloud using as much data as possible. The trained algorithm was then deployed on the sensors so that the computational effort was minimized while accuracy was still maintained. Improvements in sensor hardware as well as new algorithms helped this effort. As one of the scientists said - “The image analytics capabilities of the smaller, more affordable sensors that we’re making allow for a lot of things that were previously impossible because of price and resolution limitations”. 


Building these low-cost systems is not just something that can be deployed in disasters, but could also be used in communities and systems that have limited resources.

What our community are reading

Moonshots, Models, IoT and Machine Learning in Agriculture

  What do Google, Climate Corporation, early stage startups in farm robotics, and researchers trying to figure out how to feed the world sustainably have in common? They’re all grappling with one of the toughest challenges of working with natural systems - how do you work with data that is sparse, unevenly distributed and with systems that have so many connections and interactions with other systems?     Before the advent of cheap sensors that are connected to phones, easily accessible satellite data and drones that can fly over fields quickly and inexpensively -     scientists in companies and academia worked on developing plant and crop models that incorporated as many aspects of the farm and as much data as was available so that they could understand and predict what was likely to happen on the field. Understandably, the forecasts took some time to produce and as the models grew more complex, issues about how to estimate model parameters and the uncertainty associated with the resul
Read more

Our online community space is now live!

 Our online community space is now open to anyone who has signed up for a free or paid course on our website! In addition to everyone who signed up for our cohort-based courses, we're now expanding it to all the members of our community. If you've already signed up for any of our courses, check your email for the invitation for the space. It's where we'll get together to talk about all things data science and clean technology related, discuss the latest research, network and make connections with other professionals in the sector. It's an invitation only , no bots and no trolls allowed space - so come on over! Here's where you can check out our courses and join our community !
Read more

How much water should an email consume? Data centers and water use

  A mid-sized data center consumes around 300,000 gallons of water a day, or about as much as 1,000 U.S. households; About 20% of data centers in the United States already rely on watersheds that are under moderate to high stress from drought and other factors; Operating a data center often requires a tradeoff between water use and energy use; And in a survey of 122 data centers in the United States, only 16% or 20 utilities reported plans for managing water-related risks. As professionals working in the field, what can we do to solve this issue? One aspect is developing and using water models that can identify water risks at different scales - so that we can predict the risk to water supplies under a changing climate. A second is using machine learning to identify and optimize water use between all the stakeholders in the watershed - data centers, farmers, cities, other industries - so that biases and needs are brought out into the open and the key issues identified. A third, of cours
Read more