LocalHawk is a multidisciplinary student project run by Kongsberg Defense Systems. This year, the project consists of 12 students with different backgrounds and expertise working together towards a common goal - to create a first response system for the protection of security-critical infrastructure.
Within seven weeks, twelve students from different technological branches take part in resolving the complex technological task of developing an autonomous drone swarm that will work as a surveillance system. This system’s main functionalities are to observe, orient, detect, and act upon an intruder.
During the operation, the system shall detect - which indicates distinguishing the object from background, recognize the object, identify the object, and then decide its course of action based on observations alone.
In the summer of 2019, LocalHawk will be used as a sensor at a surveillance system created to survey a security critical area, developed by the engineers at KONGSBERG. Localhawk is one of multiple sensor systems that will provide the surveillance system with data. All the data from the different sensors will be processed with sensor fusion to provide situational awareness.
The students will experience how it is to work as an engineer in KONGSBERG. This project is integrated with other on-going projects at KDA DM, while having its own project manager and system architects. As a member of the team you will develop both technical skills and ability to cooperate and make plans in order to finalize a product by a given deadline.
Autonomy and sensor fusion
The drone swarm shall be able to observe, orient, detect and act upon an intruder. And further provide the operator with a live video stream to gain additional situational awareness. The drones shall be able to track moving intruders and avoid objects in its way. This involves developing autonomous software, sensors, object detection algorithms, power design and route planning.
Through the design, development and test phases, students are supported by highly skilled and experienced KONGSBERG engineers.
The software is based on Robot Operating System (ROS) framework and is implemented in C++ and Python. LocalHawk uses multiple of Intel Aero Ready to Fly drones, and these run on Linux Ubuntu. Neural networks run directly on the drone to identify and classify the objects moving inside the security critical area.