On 16-17 June 2026, Senior Researcher Kārlis Freivalds and Researcher Laura Leja from the Institute of Electronics and Computer Science participated in the AIMS5.0 final general assembly meeting in Madrid, Spain. The meeting brought together project partners to present the results of the project work packages, use cases, demonstrators and collaboration outcomes.
At the event, EDI presented the results of Work Package 5, Use Case 6: “Reconfigurable AI-based automated conveyor feeding by a robotic system”. The results were showcased through a poster and demo video describing an AI-based robotic system for automated bottle picking and insertion into sockets on a conveyor. The use case aims to replace manual labour with a flexible and reconfigurable robotic solution that combines camera data, bottle and socket detection, 6D pose estimation, grasp planning, socket tracking, trajectory planning and dynamic robot control.
The demonstrated solution shows an end-to-end scenario in which a robot picks bottles from an unstructured environment and places them into random sockets, adapting to dynamically detected bottle and target positions. Special attention was given to Differentiable Physics-based methods, which enable smooth and precise motion planning and allow the robot to adapt to changing conditions with low computational demands.
In addition, Kārlis Freivalds presented the results of work package 2 on AI tools, methods and algorithms for sustainable industrial processes. Laura Leja also contributed to the preparation of the work package 8 presentation, dedicated to dissemination, standardization and exploitation of the project results.
