Remotely Piloted Aircraft Systems (RPAS), commonly known as drones, are becoming widely used as an aerial data capture platform in engineering surveying and infrastructure incidents handling. Semi-autonomous RPAS performing a mission of assets inventory, damage assessment or threats detection without the involvement of human resources would be an invaluable tool. With further automatic analysis of real-time sensor data with added dimension of location against customer-specific domain system data (assets and infrastructure management, GIS, etc.), organizations may gain completely new level of knowledge and decision-making capabilities. It may sound like sci-fi, but in fact it is real business concept.
An idea for a Data Science Team presentation
A few weeks before SAP Forum Poland in Sopot, we were asked to turn our experience into a nice presentation. We decided to raise the challenge a bit higher and not only prepare additional real-time live demo but also give an access by QR code to the app to all interested participants at the same time.
The demo scenario seems simple: collect & transfer via GSM live sensors data from an RPAS during its mission over the Baltic Sea and conduct real-time analysis with SAP HANA platform. As you can imagine, the idea was quite risky – we had to take into account, inter alia, the weather (wind and rain), possibility of a failure of a drone, GSM internet connection, video-streaming over optical fiber, HANA Appliance and custom application with HERE client, network throughput, but also… BALTOPS 2015 – the biggest NATO Allies naval exercise since Cold War.
Special missions always require a special crew. It was no different this time. We wouldn’t have succeeded without the great support of our colleagues from HP who have delivered the requested HANA Appliance and Novelty RPAS who provided a world class unmanned multirotor system “Ogar”. Thank you!
To keep the presentation on the track of this year’s SAP Forum Poland motto – “Discover simple”, I’ve decided (solely responsible for custom development) to simplify the solution architecture where applicable and reasonable.
The backbone of the application was data model defined within CDS/HDBDD artifact and exposed via REST/oData service, which was also being used by drones to push data. The raw sensor data coming at regular intervals via HTTP POST request were enriched within custom insert procedure with sequence ID, drone ID (based on basic authentication header) and native ST_POINT geospatial type. Final collections of table entities and parameterized calculation views were exposed for public consumption (anonymous connection via restricted user with granted hdbrole).
From the participants’ front-end perspective, custom User Interface based on Nokia HERE v3 client was prepared. Its main task was to track flying objects and display simple server-side calculations results acquired through regular async calls to REST service collections.
XS, with all its pros and cons, was a good choice to demonstrate the simplicity since the final project (maintained as design-time objects in HANA repository) contained only of a few hundred lines of source code, graphical models and native artifacts.
Despite the fact, that the technical crew encountered a problem with the optical fiber just a minute before the start and seagulls didn’t like the new flying friend to such an extent that they were trying to knock it down, we can probably consider the show a success, that can be measured by regular access to the demo app up to 3am during the celebration night.
If you have any questions, please feel free to leave your comment below – I’ll be glad to answer it.