FOCUS ON IMADA: design and evaluation methods for data analysis of the drone mission


IMADA is the short name for the Department of Mathematics and Computer Science at the Faculty of Science, University of Southern Denmark (SDU). Our role in D4S is to develop a cloud system for managing autonomous inspection missions using drones. This system will provide a user interface for high-level inspection mission planning and monitoring. The user will be able to select an infrastructure that needs to be inspected on a 2D map and calculate a near-optimal flight route to reach it. Then, the system will send the navigation data to the drones, which will fly to inspection targets. The user will be able to monitor the movement of the drones and its sensor readings through the user interface to assure the inspection progresses as expected. The drones will deliver inspection images to the cloud system, where the user will be able to further analyse or upload them for an AI analysis.


CLOUD PLATFORM’S DEVELOPMENT: stages of the process

Currently, we are working on the development of a cloud platform to support autonomous inspection mission management. That includes mission planning, communication with the drones, and monitoring of the mission.
First, when the project had started, we analysed the requirements of the systems and designed the software architecture accordingly. The design encompasses modular units, each providing individual functionality, shaping the system in a microservice architecture. In this way, it is comparatively easy to add new features to the system and to update the old ones when needed. We also deploy the application to the cloud in a Kubernetes cluster to take advantage of its scaling properties. With Kubernetes, the system is ready to support numerous users in the future.
Meanwhile, we have proceeded with the development of services for retrieval of the infrastructure locations from open data sources, route calculation, data storage, and delivery, location estimation, and image upload for AI analysis. Currently, we are developing a service for exchanging data with the drones.


The main objective of the project is to develop a system for infrastructure inspection using autonomous drones. However, even autonomous drones need to know what their mission is. The application we are developing will provide drones with high-level navigation data to instruct them where to fly and what to inspect. Furthermore, we want to assure the safety of the mission. It is, therefore, important to be able to monitor the progress and status of the drones at every moment. The user interface, developed during our work, will provide a bridge (no pun intended) between the drones and the user. The user will be able to follow the progress in near real-time and react if an unexpected situation occurs. Last, but not least, the user will be able to visualize mission results, send them to the analysis system, and store all the data gathered through the mission. We expect that the amount of data gathered through numerous missions will help us to find patterns that could further optimize infrastructure inspection processes.

Peter Schneider-Kamp
Peter Schneider-Kamp, full professor of computer science at SDU IMADA

There are a few challenges we encountered while working on the D4S project to enable autonomous infrastructure inspection. One worth mentioning is that regulations for autonomous flights are highly constrained or even undefined. As this inspection solution is revolutionary and involves the newest technology, we experience the regulations as rather restrictive, and it takes a significant amount of time to obtain the required permissions to even be allowed to test prototypes. To overcome that challenge, we are thinking about safety and security in every step of the development process, analyzing potential threats, and implementing solutions. We have been spending extended amounts of time understanding the regulations and implementing proof-of-concept solutions to test the feasibility of our work. Another challenge is the immaturity of some technology that we can potentially employ for autonomous inspection. That is something we have been aware of from the beginning of the project and what attracted us to this topic. So instead of overcoming this challenge, we fully embrace it and put our minds into research every day. We rely on the wider research community and keep up to date with their work.


DRONES4SAFETY as a basis for future projects

As D4S is one of the first big projects for autonomous infrastructure inspections, we see it as a foundation for all future work concerning this topic. Our wish is to increase the safety of infrastructure inspectors and avoid accidents by employing autonomous drones. Furthermore, by enabling continuous infrastructure inspection we wish to prevent accidents caused by unmaintained transport infrastructure around Europe.
We are happy to have an opportunity to work on the D4S project. Every day is a new challenge, and we enjoy collaborating with our project partners. The feeling of working on an idea never implemented before, which could make our lives better in the future, is priceless.

Author :
Peter Schneider-Kamp, full professor of computer science at SDU IMADA and leader of the SDU strategic focus area Open Data Exploration, with the contribution of Lea Matlekovic, PhD fellow at University of Southern Denmark. 

Leave a Reply

Your email address will not be published. Required fields are marked *