Virtual Shepherd: Guiding Livestock Without Fences – How Geodata Makes the Alps Safer
MovingLayersCattle scattered across an entire alpine pasture, barely any mobile coverage, steep terrain – and yet the position of every animal should be known in near real time: that is the spatial starting point of Virtual Shepherd, an FFG project within the Austrian Space Applications Programme (ASAP, 2024–2026). The goal was to replace grazing fences with virtual boundaries (“fenceless farming”) and to defuse conflicts between livestock and hikers. As MovingLayers, we were responsible for the system architecture and the two demonstrator apps – and encountered three GIS challenges along the way.
Positioning without a network
In alpine terrain, connectivity is patchy. Positioning therefore combines GNSS with terrestrial LoRaWAN as well as 4G/5G; processing runs at the edge – directly at the base station – rather than in the cloud. Where no network is available at all, a Starlink link provides the return channel. This keeps the position available even where conventional tracking solutions fail.
Sensor data in real time
Hundreds of position reports per day must be structured, stored and retrievable. We connected the sensor data via a FROST server following the OGC standard SensorThings API and processed it in a cloud environment on Microsoft Azure. This way, collar, base station and app all speak the same, standardised geodata language.
Detecting anomalies spatially
Trajectories only become knowledge through analysis. Heatmaps, hexagonal grid maps and a space-time cube make movement patterns visible; anomaly detection reports deviations – for example an animal that has not moved for an unusually long time or that leaves the virtual grazing boundary. The farmer receives a notification and can check on the spot.
Two apps, two audiences
Value emerges on the device. We developed two demonstrator apps: one for farmers, managing animals and virtual fences and receiving alerts; one for hikers, who use GNSS to draw a temporary protective radius around their own position so that livestock and people stay apart. Both were released in early 2025 as iOS and Android applications.
Result
Virtual Shepherd proved the chain from collar to app in the field – including Starlink connectivity and real alpine farm operations. In January 2026, the project concluded with a final meeting in Raumberg-Gumpenstein, together with participating farms and the Austrian Alpine Pasture Association (Österreichischer Almverband).
📋 Project fact sheet — Virtual Shepherd
Funding: FFG, Austrian Space Applications Programme (ASAP), 2022 call · Duration: 2024–2026
Consortium: University of Salzburg / Z_GIS (lead), MovingLayers, TU Wien, Viehfinder (Peter Lederer)
Methodology: GNSS/LoRaWAN/satellite, edge computing, SensorThings API (FROST), GeoAI anomaly detection
MovingLayers’ role: System architecture, cloud (Azure), two demonstrator apps, field evaluation
Analysing the herd’s movement data in space
Where is the herd, what patterns do the trajectories reveal, and when does a deviation become an alert? The Virtual Shepherd IoT Monitoring System shows in detail how the monitored livestock can be analysed spatially.
