We have submitted our latest results to the ICRA 2026 Workshop on Field Robotics, presenting a prototype of Physical AI for collaborative robotics. The work demonstrates cooperative beam lifting by two mobile manipulators, controlled by AI policies trained in simulation and executed in a decentralized way on real robots at Fraunhofer Read more…
ENKI AI is developing a prototype of a collaborative multi-robot system capable of jointly lifting and stabilising a beam using embedded AI control. Two robotic units coordinate their motion in real time, continuously exchanging state information and adapting their behaviour to maintain alignment, height synchronisation and load stability. The system Read more…
Increasing productivity, improving precision placement, and enabling new collaborative construction operations that have not previously been feasible at scale — this is the objective of our work at ENKI AI. We are advancing a new approach to collaborative robotics in modular construction by integrating collaborative multi-robot scheduling, real-time network control Read more…
We are conducting early-stage research on a ROS 2-based Collaborative Robot Orchestrator (CRO), exploring new approaches to coordinating collaborative tasks among multiple robots. In this conceptual framework, the CRO investigates how a shared collaborative task can be assigned to a group of robots via wireless connectivity, including 5G and prospective Read more…
The 5G NG-RAN Simulator is a vital tool for developing AI/ML algorithms aimed at optimizing the Radio Access Network (RAN). It is designed to serve as the environment for reinforcement learning. The reinforcement learning approach in machine learning relies on the concept of an environment. The RAN Simulator functions as Read more…
ENKI AI has developed the KPM xApp, which subscribes to cell-level Key Performance Measurements (KPMs) available at the O-DU and O-CU, presenting the reported metrics in charts via the popular Grafana application. KPMs for individual cells can be displayed with 1-second granularity on the time axis, enabling detailed analysis. The Read more…
ENKI AI has developed a cutting-edge O-RAN compliant Near-Real-Time RAN Intelligent Controller (Near-RT RIC). Built entirely in Python for rapid extension and maximum flexibility to support a wide range of research directions. The Near-RT RIC, along with its xApps, serves as the central node for optimizing Radio Access Network (RAN) Read more…