The design philosophy of CyberCortex.AI is that the underlying perception and planning mechanisms of a robot's artificial brain should be the same, regardless of the robot's kinematic chain (wheeled, legged, or aerial).
AI-based perception and motion control algorithms for autonomous robots.
Optimized deep neural networks used for sensing and SLAM.
Controlling distributed teams of robots operating in complex environments.
cloud
dnn design
dnn training
data storage & retrieval
CyberCortex.AI was designed in order to create AI-based, distributed and scalable robotics applications. It provides the following features:
Our main product is the CyberCortex.AI framework, composed of the real-time CyberCortex.AI.inference operating system and its additional components. Together, these can be used to construct various real-time robotics applications based on customer tailored Artificial Intelligence components.
Based on CyberCortex.AI, we are able to provide tailored solutions for various robotics applications, ranging from autonomous driving and logistics to applications targering embedded systems, such as computer vision on the edge, robotics in agriculture, or monitoring maintainance.
Using our over 15 years know-how in the area of robotics, artificial intelligence and computer vision, we provide consulting services for companies looking to scale up their business and portofolio using AI software.
Grigorescu, S. M., & Zaha, M. (2023). Method and AI operating system for robotics and complex automation. European Patent No. EP4290822A1. European Patent Office. https://patents.google.com/patent/EP4290822A1/en
Grigorescu, S., & Zaha, M. (2025). CyberCortex.AI: An AI-based operating system for autonomous robotics and complex automation. Journal of Field Robotics, 42(2), 474-492. https://doi.org/10.1002/rob.22426
@article{CyberCortex_AI,
author = {Grigorescu, Sorin and Zaha, Mihai},
title = {CyberCortex.AI: An AI-based operating system for autonomous robotics and complex automation},
journal = {Journal of Field Robotics},
volume = {42},
number = {2},
pages = {474-492},
doi = {https://doi.org/10.1002/rob.22426},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/rob.22426},
year = {2025}
}