iRonCub3 becomes the first humanoid robot to fly using jet engines
AI-powered neural networks enable real-time aerodynamic flight control
Robot integrates multibody dynamics with thrust and balance strategies
iRonCub3 becomes the world’s first humanoid robot to fly using jets and AI controls.
Photo Credit: IIT-Istituto Italiano di Tecnologia
A jet-powered human robot has taken to the skies, and this time it's managing to stay in the air and even move around mid-flight, thanks to some nifty AI and heavy integration. The iRonCub3, developed by the Italian Institute of Technology (IIT), managed to fly in the air at 50 cm off the ground during testing – a massive advance in the world of robotics. Engineered to operate in the real world, the 70-kilogram robot is powered by four microjet turbines and advanced neural controllers. This prototype of the testbed system can stably hover in gusty wind conditions. The study, inspired by two years of design, experimentation, and simulation, seeks to extend robotics into extreme, unstructured terrains with aerial capabilities.
As per a Communications Engineering report, the project is spearheaded at IIT's Artificial and Mechanical Intelligence (AMI) Lab in Genoa by Daniele Pucci, along with Milan Polytechnic and Stanford University. The group designed a neural network-based real-time aero analysis integrated AI system. Such systems, learnt based on both experimental and simulation data, enable the robot to maintain stability in takeoff and during dynamic flight. Additional contributions include thermodynamic assessments, wind tunnel tests, and aerodynamic modelling of the robot's human-like multibody form.
Unlike traditional drones, the iRonCub3's limbs are elongated and movable, which forced researchers to develop models that could account for thrust, mass, and motion. Meanwhile, titanium spine reinforcements and heat-resistant components mean it can cope with exhaust from the turbines at a blistering 800°C, while precision sensors, integrated jet control, and body re-engineering allow for secure and precise takeoff and flight. Neural models manage rapidly shifting airflow and posture changes caused by limb motion.
Engineers used a co-design approach to determine optimal turbine placement and cooling strategies. The real-time AI feedback loop allows the robot to navigate flight with human-like balance, despite aerodynamic complexity. Hovering trials at IIT's facility have been successful; tests at Genoa Airport are next in line, complying with safety regulations.
The developed humanoid aerial robot can be used for disaster search and rescue, dangerous field inspection, and autonomous home exploration. Combining flight with human-like mobility, iRonCub3 marks a pioneering breakthrough in robotics and a potential model for agile machines designed to work in places where humans and traditional drones can't easily reach.
MIT Develops Low-Resource AI System to Control Soft Robots with Just One Image14 July 2025
Search
![[Sponsored] Students Try the Galaxy Tab S10FE for the First Time - Real Reactions, Powerful Features](https://c.ndtvimg.com/2025-07/54d9ikeg_galaxy_160x120_18_July_25.jpg?downsize=180:*)
