In norse mythology Gjallarhorn is Heimdall’s horn, which signals the beginning of Ragnarök and be heard around the world.  

While security threats are becoming increasingly complex and multifaceted, the mental capacity of operators monitoring sensors and alarms is quickly reaching a maximum. At the same time newer and better technology is developed and shoe-horned into our old systems. This thesis explores how machine learning can be implemented in surveillance operations to reduce alarm fatigue and enhance the human operators’ capabilities by supporting their situation awareness.  

By analyzing both external actors in the domain and collaborators from the Norwegian Armed Forces and the Norwegian Home Guard, the current and desired future landscape has been mapped out. Key deliverables include a communicable vision of moving from alarm-based monitoring to event-based monitoring, user journeys for what this change allows, and prototypes for short-term and long-term value.