Researchers from Shibaura Institute of Technology, Waseda University, and Fujitsu have developed a novel way to make robots move smoothly and efficiently using quantum computing.
Usually, when a robot moves, its computer has to calculate how each of its joints should bend so that its hand or foot ends up in the right spot. This process, known as inverse kinematics, is extremely difficult for robots like humanoids because there are countless possible combinations.
Traditional computers must do many trial-and-error calculations, which takes time and computing power.
A quantum solution
The team’s new approach uses qubits to represent the position and orientation of each part of the robot. Even more importantly, they use quantum entanglement – a special feature of quantum mechanics where particles are linked so that the movement of one affects the other.
This mirrors how real robot joints work, where moving one joint affects the connected joints.
Another important aspect of this research is the hybrid approach, combining classical and quantum computing. While the forward kinematics – calculating where the robot’s hand or foot ends up given certain joint angles – is handled by quantum circuits, the inverse kinematics step is still performed on classical computers.
This division of labor allows the system to take advantage of quantum speed-ups while maintaining stability through conventional methods.
Faster, more accurate calculations
By doing this, the researchers were able to reduce the number of calculations needed. Tests on Fujitsu’s quantum simulator showed that the method cut errors by up to 43% compared to classical methods, and it ran faster. They even confirmed the results using a 64-qubit quantum computer developed with RIKEN.
In one test, they tried to calculate the movements of a full-body robot with 17 joints(similar to a human). Normally, this would take an unrealistic amount of computing power and approximately 30 minutes to complete.
Why this matters for future robots
This breakthrough matters because future robots – especially humanoid robots that work closely with humans – must move fluidly, respond quickly, and handle complex environments in real time.
Current methods often simplify the model (for example, reducing the number of joints in the calculation to 7 instead of 17), making movements stiff and less lifelike.
With the new quantum-based method, smoother and more realistic robot movement could become possible. And the best part? It can already work on today’s “NISQ” quantum computers – machines that are not yet perfect, but usable for certain problems.
In the long run, this technology could help robots with real-time control, obstacle avoidance, multi-joint manipulators, and energy optimization tasks.
Looking ahead
The researchers also believe the approach could get even better if combined with advanced quantum algorithms, such as the quantum Fourier transform, which might further speed up calculations.
By blending quantum computing with robotics, the team has taken a big step toward creating the next generation of intelligent, human-like robots.
Takuya Otani from the Shibaura Institute of Technology and Atsuo Takanishi from Waseda University collaborated on this research, alongside Nobuyuki Hara, Yutaka Takita, and Koichi Kimura from Fujitsu Limited.