More efficient afterharvest logistics in the tomato greenhouse
In order to further automate logistical processes in greenhouse horticulture, an Autonomous Mobile Platform (AMP) is being tested. This platform must navigate independently within the greenhouse environment and perform path changes without human intervention. The project validates whether the AMP is able to detect variations in the subsoil and rail systems and anticipate on this.
Innovation package, use case and type test
Glastuinbouw
After Harvest
Status: in progress
Technical functionality
Broad knowledge question
How can an autonomous platform navigate safely and accurately?
The AMP is equipped with cameras and an AI model to detect and respond to variations in concrete smoothness and tube rail structures. This project investigates how the platform can move autonomously over the tube rail systems, detect obstacles and perform a correct path change.
Approach
Field tests within Tomatoworld
The AMP is being tested at a greenhouse location at Tomatoworld. During the testing phase, four crucial functions are evaluated:
Path switching between different rail systems.
Detection of whether the platform is correctly positioned on the rails.
Behavior at the end of a rail system.
Obstacle detection and response capability.
These tests aim to demonstrate whether the system is robust enough for practical use in a dynamic greenhouse environment.
Goal
Autonomous navigation without human intervention
The ultimate goal is to validate that the AMP can independently navigate from one pipe system to another without manual adjustments or external intervention. This would be an important step toward further automating afterharvest logistics processes within greenhouse horticulture.
Result and reflection
Successful implementation and lessons learned
The test focuses on whether the AMP can adequately adapt and function within the specific variations at Tomatoworld.
Successes:
The AMP can detect the rails and navigate correctly within the test area.
Initial tests show the platform is capable of performing path switches without manual correction.
Lessons learned:
Further fine-tuning of the AI model is needed to better compensate for variations in pipe rail structures.
Obstacle detection and response speed require optimization to ensure stable operation in practical environments.