Breaking new ground in robotic mobility, engineers from the California Institute of Technology have unveiled X1, a pioneering machine capable of walking on two legs, soaring through the air like a drone, and navigating terrain on wheels—all integrated into one sophisticated system. This project, a collaboration between Caltech, the Technology Innovation Institute (TII) in Abu Dhabi, and Northeastern University, represents a significant advancement in hybrid robotic technology.
X1 combines a humanoid robot with a transformable aerial-ground drone, seamlessly integrating walking, flying, and rolling capabilities. It can autonomously climb stairs, move across varied environments, deploy a drone from its back, and rendezvous with it remotely, showcasing unprecedented multifunctional mobility.
The ambition behind X1 extends beyond novelty; it aims to equip future robots with the agility required for disaster relief, urban exploration, and rapidly changing environments where adaptability is crucial. While most robots today are confined to a single mode of movement, X1 tackles this limitation head-on.
“We wanted to take the strengths of each system—walking, flying, driving—and combine them in a way that avoids their individual weaknesses,” said Aaron Ames, director of Caltech’s Center for Autonomous Systems and Technologies (CAST).
The Humanoid-Air-Ground Robot With Real-World Applications
The foundation of the X1 system is a customized Unitree G1 humanoid robot, enhanced by Caltech’s team to autonomously carry and release the bespoke M4 robot, developed at Professor Mory Gharib’s CAST laboratory. The M4 uniquely functions as both a quadcopter and a terrestrial vehicle, employing its round propeller guards as wheels when navigating on land.
During a recent live showcase, the G1 robot started within Caltech’s Gates–Thomas Laboratory, progressing through indoor corridors, ascending stairs, and handling outdoor surroundings before deploying the M4 drone. The airborne M4 maneuvered past obstacles, landed, and seamlessly transitioned to rolling mode to finish the designated course.
This exercise simulated an urgent rescue scenario, testing X1’s performance over diverse terrains. Notably, the robot managed all transitions—walking, flying, and driving—completely on its own without external control during the mission.
Essential elements of the platform include:
- Unitree G1: a nimble humanoid platform adapted for autonomous exploration and drone launching.
- M4 Robot: an agile drone engineered for rapid switching between aerial and ground operation modes.
- Integrated AI and sensors: systems enabling environmental detection, obstacle navigation, and autonomous mission planning.
Designed for Environmental Adaptability, Not Mere Emulation
What sets X1 apart is its capability to learn and adapt dynamically, going beyond simple preprogrammed actions. Unlike many humanoids that depend on replicated human movement data, X1’s creators employed physics-driven modeling combined with machine learning so the robot can independently develop effective movement techniques tailored to real-time conditions.
“The robot learns to walk as the physics dictate,” Ames explained in Caltech’s official statement. “It can walk on different terrain types, it can climb stairs, and—importantly—it can do so while carrying the M4 on its back.”
Such autonomous decision-making is critical for robots operating in hazardous or unpredictable settings. Whether navigating disaster zones, smoke-filled structures, or rough terrains, a robot’s capacity to select the safest and most suitable mode of travel could ultimately influence mission success.
An International Collaborative Effort
This endeavor taps into global expertise: Caltech’s Gharib lab contributed deep knowledge on biomimetic flight and shape-shifting aerial vehicles, while Ames’ group specializes in control systems and bipedal locomotion. The Technology Innovation Institute added sophisticated autonomous sensing and implemented the secure Saluki flight controller for the latest M4 iteration.
“The challenge is how to bring different robots to work together so they become one system,” said Gharib, who has led CAST since its founding.
TII engineers also developed a sensor fusion architecture that integrates lidar, cameras, and range finders for X1, enabling comprehensive, continuous environmental awareness. This allows the robotic platform to independently decide whether to walk, fly, or roll based on its surroundings.
Research contributions from Northeastern University focused on the drone’s morphing technology, ensuring smooth transitions between aerial and terrestrial modes without mechanical failure or lag.
Future Prospects
While still in research stages, X1’s capabilities suggest promising uses across various fields. Its versatility, autonomous intelligence, and swift adaptability position it as an ideal candidate for:
- Emergency response and search operations
- Military reconnaissance missions
- Automated supply delivery in rugged environments
- Inspection and maintenance of complex infrastructure
“We’re thinking about safety-critical control, making sure we can trust our systems,” Ames noted. “These problems are really big. But we’re taking them on in a substantial and concerted way.”
Innovations like X1 demonstrate the evolution from isolated robots to interconnected, multimodal machines capable of adapting fluidly within their environments. With ongoing improvements in safety, reliability, and practicality, platforms like X1 could soon become indispensable allies in urban navigation and critical rescue scenarios—mastering not just movement, but transformation.
- Categories:
- News
0 comments
Sign in to Comment