Friday 21 March 2025
A team of researchers has made significant strides in developing a method for robots to autonomously discover the properties of objects they’re manipulating, without relying on expensive sensors or communication between machines.
The approach, described in a recent paper, involves using twists, twist derivatives, and wrenches measured at each robot’s grasp frame to estimate the transformation matrices between those frames, as well as the location of the object’s center of mass and its inertia matrix. In other words, the robots are able to figure out where they’re attached to the object, how heavy it is, and how it moves.
The team used a group of Omnid robots, designed for human-robot collaboration, to test their method in a series of experiments. The results showed that the robots were able to accurately estimate the object’s properties, with errors typically less than 3 degrees for rotation matrices and less than 4% for position vectors.
This development has significant implications for robotics research, particularly in areas like multi-robot manipulation, where multiple robots work together to move or manipulate objects. Currently, this type of collaboration often requires complex communication networks and expensive sensors. By allowing robots to autonomously discover the properties of objects, researchers can simplify these systems and make them more practical for real-world applications.
The method also has potential uses in areas like search and rescue, where robots may need to work together to move debris or manipulate heavy objects. In these scenarios, being able to quickly and accurately determine the properties of an object could be a matter of life and death.
One of the key challenges the team faced was dealing with the noise and uncertainty inherent in the data collected by the robots’ sensors. To address this, they employed advanced filtering techniques and statistical methods to improve the accuracy of their estimates.
The researchers also explored ways to extend their method to objects that are not rigid or have complex shapes. While these scenarios present additional challenges, the team’s approach provides a solid foundation for tackling them in future research.
Overall, this work represents an important step forward in the development of autonomous robotic systems capable of collaborating with humans and each other to accomplish complex tasks. As robotics continues to play an increasingly important role in our daily lives, innovations like this will be crucial for enabling robots to safely and effectively interact with their environment.
Cite this article: “Robots Learn to Discover Objects Properties on Their Own”, The Science Archive, 2025.
Robotics, Autonomous Systems, Object Manipulation, Sensor-Free, Robotic Collaboration, Multi-Robot Manipulation, Search And Rescue, Filtering Techniques, Statistical Methods, Robotic Grasping.
