Robotics and quantum computing, two of the most advanced technological fields of our time, have recently intersected at a pivotal point. Researchers from the Shibaura Institute of Technology, Waseda University, and Fujitsu Limited have announced the development of an innovative method to optimize robot postures using quantum techniques, achieving reductions in errors and calculations in one of modern robotics’ biggest challenges: inverse kinematics.
The challenge: how to achieve natural movements in humanoid robots
A humanoid robot, or a robotic arm with multiple joints, must constantly calculate the angles to adopt in each joint to reach a final position—for example, picking an object from the ground without losing balance. This operation, known as inverse kinematics, involves solving thousands of possible combinations within seconds.
In a complete model with 17 joints, equivalent to the main joints of a human body, the computational load is so high that classical methods resort to simplifications, such as reducing the model to 7 joints. However, these approximations limit the smoothness and complexity of movement, making current robots still appear clumsy compared to humans.
The proposal: representing postures with qubits
The new method introduces a quantum approach to the problem. According to the team, each orientation and position of the robot’s parts (the so-called links) is represented using qubits, the fundamental units of information in a quantum computer.
The process operates in a hybrid manner:
- The calculation of forward kinematics (determining where the hand ends up from the joint angles) is performed with quantum circuits.
- The calculation of inverse kinematics (solving for the angles based on the desired position) is executed on classical computers.
The key lies in quantum entanglement: this phenomenon allows the movements of one joint to automatically influence the subsequent ones, naturally replicating the hierarchical dependence of the human skeleton.
Results: fewer calculations, greater accuracy
Tests conducted with the Fujitsu quantum simulator achieved a 43% reduction in errors compared to conventional methods, while also using less computational resources.
In an additional experiment with the 64-qubit quantum computer developed by RIKEN and Fujitsu, it was confirmed that entanglement significantly improved convergence speed and calculation accuracy.
In a test case with a complete model of 17 joints, the system was able to compute movements in about 30 minutes, a figure unthinkable just a few years ago.
Implications: moving toward more human-like robots
This breakthrough, still in experimental stages, opens the door to robots capable of executing smoother, more precise, and more complex movements, even in scenarios involving human interaction.
Immediate applications include:
- Humanoid robots for assistance at home or in healthcare settings.
- Industrial manipulators with greater precision in factories and assembly lines.
- Rescue robots capable of adapting to uneven terrains or unpredictable obstacles.
- Energy optimization, adjusting postures to minimize effort and energy consumption.
Furthermore, since the technique requires a reduced number of qubits, it could be implemented even on current NISQ (Noisy Intermediate-Scale Quantum) computers, accelerating the transition from theory to real-world applications.
A partnership shaping the future
This project results from collaboration among three leading Japanese institutions:
- The Shibaura Institute of Technology, led by Associate Professor Takuya Otani of the Human-Robot Systems Laboratory.
- Waseda University, with Professor Atsuo Takanishi, a pioneer in humanoid robotics.
- Fujitsu Limited, represented by Nobuyuki Hara, Yutaka Takita, and Koichi Kimura, responsible for integrating quantum capacity into robot simulation and control.
The work has been published under the title Quantum computation for robot posture optimization, marking a first step toward the practical use of quantum computing applied to real-time robotics.
Convergence of trends: quantum, AI, and robotics
This advancement does not stand alone. Japan, through projects like FugakuNEXT—the supercomputer developed by RIKEN, Fujitsu, and NVIDIA—aims to integrate artificial intelligence, high-performance computing, and quantum technologies within a single platform.
The vision is clear: to create robots that not only mimic human appearance but can also learn, adapt, and respond with the same agility as humans.
Conclusion
The announcement by Fujitsu, Waseda, and Shibaura marks a milestone in robotics history. For the first time, quantum computing has been successfully applied to one of the most complex problems of humanoid movement: inverse kinematics.
Although there is still a journey ahead to achieve real-time calculations, the results foreshadow a future where robots will move with unprecedented smoothness, precision, and energy efficiency.
According to the researchers, this is “a breakthrough that could transform how humans and machines coexist in everyday environments.”
Frequently Asked Questions (FAQ)
1. What is inverse kinematics in robotics?
It is the mathematical calculation that determines the angles of each joint to reach a desired final position—for example, grabbing an object with the hand.
2. Why does quantum computing improve these calculations?
Because qubits and entanglement allow the representation of multiple states simultaneously, reducing the number of calculations needed and increasing accuracy.
3. Can this be applied to commercial robots already?
So far, the technique has been tested in simulators and experimental quantum computers, but it is feasible on current NISQ systems. The challenge remains to perform calculations at real-time speeds.
4. What practical applications could it have in the short term?
Assistive robots, industrial robotic arms, rescue systems, and advanced mobility technologies in humanoid robots.
via: global.fujitsu