This demo aims at showcasing different technologies in an intelligent agriculture context:
- AI model training and serving for disease recognition in crops
- 5G slices for two way communications between IoT/Field "devices"
- Edge computing at Telco location (MEC)
- Path optimization using Optapy and a PathFinding algorithm
- Automated model updates
Defining the obstacles
Finding options
Optimal path
Optimization with OptaPy
Some crops are ill
Everything healthy again
sequenceDiagram
participant drone as Drone
participant classification as Classification API
participant model as Model Serving API
participant tractor as Tractor
participant path as Path Services API
rect rgb(255,255,255)
note right of drone: Image classification
drone ->> classification: Status of this field? (/classify)
classification ->> model: Prediction for this field? (/prediction)
note over model: Make prediction and score
model ->> classification: Here are the results
alt Field must be treated
classification ->> path: Add this field to destinations (/destination)
end
classification ->> drone: Update field according to status
end
rect rgb(255,255,255)
note right of tractor: Tractor route computation (multiple destinations)
loop Every 10s if tractor Route is empty
tractor ->> path: Fields to treat? (/routefinder)
end
alt Destinations is empty
path ->> tractor: No, you're clear
else Destinations has entries
note over path: Find best route within tractor capacity
path ->> tractor: Yes, here is your full route
end
end
rect rgb(255,255,255)
note right of tractor: Tractor movement
tractor ->> path: What's the path to go there? (/pathfinder)
path ->> tractor: Here is the path to follow
tractor ->> path: I have treated this field
note over path: Remove field from destinations
path ->> tractor: Ack, go to next destination
end