Our comprehensive solutions help you overcome the challenges of developing autonomous driving in agriculture.
Realistic Environment
Create realistic environment sensor data for your agriculture application.
Free Ground Truth Data
Test your perception algorithms with free ground truth data.
Fleet Tests for all Vehicle Variants
Simulate fleets of any autonomous machinery on same field.
Real-Time Testing
Performant simulation enables HIL testing in real time.
The Challenge
The general challenge in testing autonomous agricultural machinery is the difficult development process with its complicated algorithms. The harvesting process is not reproducible as the crops are no longer on the field after they have interacted with the agricultural machine. In addition, the replay of the recorded sensor data is limited, e.g., because you only harvest your crops once.
So the question is: How can these limitations be overcome?
The harvester interacts with the crop in real time and these changes in the environment are updated live in all sensors.
This lets you test autonomous agricultural machinery in HIL and SIL simulation.
Difficult Development Process and Limited Data Replay of Recorded Sensor Data
- Testing in real life is expensive and might even be dangerous
- Not reproducible since crops are no longer on the field after they interacted with a farming machine
- Limited amount of test data available since you harvest crops only once
- Usual test mechanics are to record real data when field is harvested
- Corner cases like a deer in a field might not occur when you record data
- Trajectories of recorded data cannot be altered by your algorithms
The Solution
There is a way to overcome the challenges in the development process of autonomous driving in agriculture.
In simulation, you are not tied to the seasons, and with and end-to-end approach you can test perception algorithms, planning, and actuators completely in a closed loop.
Simulation Solution
- Integrate custom farming machines into simulation
- Configurable environment to virtually plant the crops you need for your scenario
- Ground truth information to test and train your AI
- Fully interactive environment influences all environment sensors
