Introducing X1: The First Multi-Robot Team Combining Humanoid and Drone Capabilities for Rescue Operations!
Imagine a scene on a sunny day in Pasadena, where a humanoid robot strides confidently across the campus of the California Institute of Technology, carrying another robot on its back. This dynamic duo, known as X1, is specifically engineered for missions that are far too perilous or disordered for human intervention.
Over the course of three years, a collaboration between engineers at Caltech and Abu Dhabi’s Technology Innovation Institute (TII) has brought these piggyback robots to life, transforming them into an integrated multirobot system. This terminology refers to a meticulously synchronized team, with each robot focusing on the tasks it performs best.
How Does the X1 Robot Duo Operate?
The project was spearheaded by Aaron Ames, a mechanical engineer who also serves as the director of Caltech’s Center for Autonomous Systems and Technologies (CAST). His research is centered around autonomous robots—machines that can plan and navigate without direct human instructions—while ensuring their balance remains intact even in challenging environments.
At the heart of the X1 unit lies a modified Unitree G1 humanoid robot. This advanced machine walks on two legs and is capable of carrying substantial equipment. On its back sits a morphing robot, a unique device that adjusts its shape according to the mission's requirements.
Beyond the capabilities of the humanoid and the M4 drone, this project brings in experts from both Abu Dhabi and Boston to ensure seamless integration among all components. At the Technology Innovation Institute, engineers are focused on enhancing secure computing systems that protect the robot's hardware, while researchers from Northeastern University fine-tune the mechanisms that enable the M4 drone to change its configuration.
In a recent report, the collaborative team highlighted how the M4 drone is equipped with cameras, lidar, and range finders, which allow it to perceive its environment effectively. These sophisticated sensors enable the duo to locate themselves without needing joystick control, facilitating route planning through congested terrains.
X1: A Robot Team Designed for Rescue Missions
During a demonstration, the X1 system began its journey inside a lab, navigated through a library, and ultimately reached a high outdoor area adjacent to Caltech Hall. This pathway presented the humanoid with narrow hallways, doorways, and steps to conquer before it could deploy its aerial companion.
Upon arriving at an open area, the humanoid bent down to allow the M4 robot to take off in drone mode. After successfully flying, M4 transformed its rotors into wheels, drove towards a pond on campus, and then soared over the water to conclude the mission.
The test followed a rescue operation script, showcasing X1 as a rapid first responder capable of surveying a scene prior to the arrival of human rescue teams. "Currently, we have robots that can fly, drive, and walk," explained Aaron Ames, underscoring the diverse capabilities of the technology.
The Power of Versatility: Walking, Driving, and Flying
Traditional single-purpose robots often encounter difficulties when faced with rough terrain, steep staircases, or situations that require sudden elevation changes. In the M4 study, researchers demonstrated that this robot can roll, fly, crouch, and maintain balance over obstacles that simpler machines would find insurmountable. This adaptability is often referred to as locomotion plasticity—the ability of a robot to alter its movement style in response to changing conditions.
M4 was designed with this flexibility in mind and scaled appropriately to carry necessary computers and sensors for truly autonomous operations. Most humanoid robots still depend on pre-recorded human movements to perform tasks like walking and climbing, which limits their effectiveness in unfamiliar scenarios. Ames’s team merges physics-based models with machine learning techniques—methods that help decipher patterns in data—allowing X1 to adapt to shifting terrain conditions.
X1 utilizes this adaptability by enabling the humanoid to traverse long distances on foot, while the M4 drone conserves energy for short bursts of driving and flight. "The challenge lies in orchestrating different robots to function as one cohesive system," noted Mory Gharib, a professor at Caltech.
From Lab Demonstrations to Real-World Emergencies
Currently, X1 operates using pre-programmed missions, but the team is actively developing safety-critical control systems that ensure safe behavior, even if sensors malfunction. This involves not only proving that the robots can reach their destinations but also ensuring they avoid hazardous actions in unpredictable conditions.
Before deploying machines like X1 in disaster-stricken areas, the team aims to gain the trust of regulators and the public. This requires demonstrating prolonged operational capability without mishaps, clear auditing methods for decision-making processes, and emergency protocols that humans can activate if necessary.
At the Technology Innovation Institute, engineers are developing secure onboard computers and flight controllers to empower the M4 drone to make swift decisions autonomously. Such capabilities are crucial in scenarios where communication networks fail, requiring the robot to navigate independently rather than relying on human pilots.
If systems like X1 continue to advance, they might revolutionize how we respond to emergencies, scouting damaged buildings or delivering supplies through flooded streets without risking human lives. The combination of legs, wheels, and rotors hints at a future where first responders operate autonomously, determining when to walk, drive, or take flight.
