Foundation AGI model for Industrial Robots

Projektinformationen

HYPER

ID Finanzhilfevereinbarung: 101218436

DOI

10.3030/101218436

EK-Unterschriftsdatum 13 Juni 2025

Startdatum 1 Juli 2025

Enddatum 31 Dezember 2026

Finanziert unter

• The European Innovation Council (EIC)

Gesamtkosten

Keine Daten

EU-Beitrag

€ 2 389 165,64

Koordiniert durch

Superintelligence Computing Systems SICSAI AB
Sweden

Periodic Reporting for period 1 - HYPER (Foundation AGI model for Industrial Robots)

Berichtszeitraum: 2025-07-01 bis 2026-03-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

The project objective is to deliver Artificial General Intelligence (AGI) foundation model for robots, a physical AGI equipped with the capability to structure information into a world map to match humanlike learning, “learning like a child” and enable automation of unstructured, unpredictable and complex environment with high variability where human labour is the only viable option today. Programmed industrial robots are reliable but rigid and cannot handle variability. Today’s AI adds flexibility but sacrifices reliability and scalability. Hybrid workarounds (legacy code patches) deployed in industrial AI robotics today don't scale. Our AI/AGI offers flexibility without cannibalizing on scalability and robustness.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

We have built a first version of our end-to-end AI/AGI robot brain that does planning and reasoning in many steps in a stable way. Our AI technology is not an LLM and does not use reinforcement learning. It learns like a child and has continual learning from feedback based on a structured world model, a prerequisite for true intelligence as described by the Turing prize Winner Yann LeCun and Nobel Prize Laureate Geoffrey Hinton.

We have benchmarked the performance of our A/AGII system with ACT, the Action Chunking Transformer (https://arxiv.org/abs/2304.13705) developed at Stanford in a cooperation with Google, used in the Aloha robot (Mobile ALOHA, Stanford Robotics Center) We measured the success rate (%) in a robot task where the robot arm and gripper interacted with a cube performing cube transfer. We demonstrated the lowest loss and fastest convergence, outperforming the baseline (ACT) with 80-100% success rate compared to 30-65%, and with 10-20 times lower loss depending on subtask and parameter setting.

We have successfully delivered an API that enables AI integration and control of industrial robots from the large robot OEMs like KUKA Robotics, ABB, Universal Robots. Our AI controls the robot all the way down to the joint angles, learning and executing trajectory, speed and acceleration and gripping force (not only picking point), which gives a higher level of flexibility in terms of handling objects of different sizes, weights and fragility.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

Our AI/AGI is beyond state of the art in AI robotics with a quantum leap in capabilities and performance delivering a union of scalability, reliability and flexibility. It has the potential to disrupt automation at its core, widening the scope and scale of AI robot automation in warehouse logistics and manufacturing industry. The universal AI/AGI robot brain can control any type of robot independent of size, shape and function, and is flexible without compromising scalability and robustness, and can automate complex tasks in unstructured environment where there are no solutions available today (80% of warehouses have zero automation and 72% of all industry assembly is still manual). It is a giant leap towards physical AGI and superintelligence. The scalability is excellent due to good generalization abilities. One universal AI controls all the different robots and learns all different tasks. This is possible since all the robots live in the same world with the same types of objects and physical laws and hence they all share one and the same world model. It is a strategic AI/AGI technology for European sovereignty and industry competitiveness.