This post is also available in: עברית (Hebrew)
A team of researchers from the Beijing Institute of Technology developed a new method to control robots that can assemble satellites in space. Their inspiration came from a surprising source- the human arm.
Robots in space operations need to interact with objects in complex and dynamic environments, but traditional robot control methods have limitations in adapting to diverse and uncertain situations, and are prone to vibration that can cause assembly failure. In order to overcome these challenges, the researchers proposed a human-like variable admittance control method based on the variable damping characteristics of the human arm.
According to Interesting Engineering, “damping” is the process of reducing the amplitude of vibrations by dissipating energy- it is essential for preventing excessive contact force from damaging the objects during assembly. The human arm can flexibly adjust its damping to perform various tasks safely and stably (the human arm reduces its damping to avoid breaking a fragile thing, while increasing its damping to exert more force when pushing a heavy object).
The researchers mimicked this feature by designing a controller that can change the robot’s damping according to the contact conditions and the assembly requirements, which can also compensate for external disturbances and environmental uncertainties.
In order to test their method, the researchers built a dynamic data acquisition platform that captured human arm motion during assembly tasks: A force sensor measured the contact force between the human hand and the assembly parts, while a motion capture system obtained the end velocity of the human arm. The researchers then analyzed the data and summarized the dynamic characteristics of the human component, as well as three contact patterns for satellite assembly: sliding contact, impact contact, and stable contact.
The researchers then applied their method to a robot that can assemble satellite components in space, and conducted simulations and experiments to evaluate the performance of their method under different scenarios. They then compared it with other compliance control methods and found that their approach could effectively improve the safety, robustness, and adaptability of robot space assembly.
In conclusion, developing control strategies that emulate human-like behavior can significantly enhance the adaptability, precision, and controllability of robots that perform assembly and maintenance tasks in space. Nevertheless, there is still more research necessary to enable robots to accomplish flexible assembly tasks comparable to those performed by humans.