In the first of a two-part article we will discuss how drones are being combined with IBM cloud services to develop an unmanned aerial vehicle to aid in search and rescue.
Drones are a remarkable piece of technology, and they’ve never been so topical. From the Civil Aviation Authority granting Amazon special permission to test its drones for package delivery in designated air spaces, to Rwanda employing the technology to get blood, vaccines and other medical supplies out to remote health centres, and drone racing captivating an online audience of 75 million people, it would seem drones have set up camp in the news headlines.
At the Imperial College London, a team of students saw a drone’s potential to aid in search and rescue, and worked with the IBM Hursley Innovation Centre to create a proof-of-concept. Often in search and rescue scenarios, there’s large, or difficult terrain to search, such as along the coast, across fields or up a cliff, or it’s dangerous to the first responders, for example, after a train crash or during flooding. Using a drone would enable rescue teams to cover wide areas and assess the situation from a safe distance before deploying resources direct to what may be a hazardous environment incident.
However, the challenge with current drone technology is that it needs to be controlled manually, which requires a trained operator. What the students wanted to test was whether they could develop drone technology using IBM cloud services to create an unmanned aerial vehicle that could be used for search and rescue.
Building the server
Working alongside the students at the IBM Innovation Centre at Hursley, we recommended using a NodeJS implementation hosted by IBM Bluemix. A cloud platform-as-a-service (PaaS), IBM Bluemix supports several programming languages and services as well as integrated DevOps to build, run, deploy and manage applications on the cloud.
The solution used two communication methods:
- MQ Telemetry Transport (MQTT Broker): specialised for small data and sensing.
- HTTPS (REST API): for large files and integration with any other component, such as the mobile app.
All the data captured by the drone was sent back to a central control system, to provide cognitive insight into the situation as it unfolded and enable the team to monitor what was happening in real-time.
This dashboard enables the user to see through the ‘eyes’ of the drone. You can see the image being fed to the camera on the bottom left, above this the drone positions an icon on the map to indicate what it has seen, and finally on the right you can see what the drone thinks it is seeing along with a cognitive confidence score. In addition, the team developed a mobile Android app where the user could monitor what the drone had determined was in the local vicinity
To validate the solution, the team repeatedly tested it end-to-end against pre-defined scenarios – it worked! Now we know that an unmanned aerial vehicle is possible, and that drones are viable technology to aid in search and rescue operations. The next steps in bringing the solution to market will involve increasing the system's robustness and durability, and longer-term, to make it more accurate, faster to respond to the needs of first responders.
In part two of this article we will look at real-life applications for the cognitive drone.