Humanoid robots could soon help humans perform a variety of tasks. Many of the tasks that these robots are designed to perform involve lifting objects of different shapes, weights, and sizes.
While many humanoid robots developed to date are able to lift small and light objects, lifting large or heavy objects has often proved more difficult. In fact, if an object is too big or heavy, a robot could end up breaking it.
With this in mind, researchers at Johns Hopkins University and the National University of Singapore (NUS) have recently developed a technique that allows robots to determine whether or not they can lift a heavy box. with unknown physical properties. This technique, presented in a pre-published paper on the archive, could allow the development of robots that can lift objects more efficiently, reducing the risk that they will lift things they cannot carry.
“We were particularly interested in how a humanoid robot could reason about the feasibility of lifting a box with unknown physical parameters,” said Yuanfeng Han, one of the researchers who conducted the study. , for TechXplore. Ã¢ â¬ ÅTo perform such a complex task, the robot usually needs to first identify the physical parameters of the box, then generate a safe trajectory of the whole body. It is stable to lift the box â.
The process by which a robot generates motion trajectories that allow it to lift objects can be computer-intensive. In fact, humanoid robots usually have a large amount of degrees of freedom, and the movement that their body must make to lift an object should meet several different constraints. This means that if a box is too heavy or its center of mass is too far from the robot, the robot will most likely be unable to complete this move.
“Think of us humans as we try to think if we can lift a heavy object, such as a dumbbell,” Han explained. Ã¢ â¬ ÅFirst we interact with the dumbbell to get a certain feel of the object. Then, based on our previous experience, we somehow know if it's too hard to lift or not. Similarly, our method begins by constructing a trajectory table, which saves different valid lifting movements for the robot, corresponding to a range of physical parameters of the box using simulations. Then the robot considers this table as the knowledge of his previous experience â.
The technique developed by Han, in collaboration with his colleague Ruixin Li and his supervisor Gregory S. Chirikjian (professor and head of the Department of Mechanical Engineering at NUS) allows a robot to obtain an intelligence. Understanding the inertia parameters of a box, for example, after briefly interacting with it. Next, the robot looks back at the trajectory table generated by the method and checks to see if it includes a lifting movement that will allow it to lift a box with these estimated parameters.
If this movement or trajectory exists, then lifting the box is considered feasible and the robot can complete the task immediately. If it does not exist, then the robot considers the load beyond its capacity.
In essence, the table of trajectories that our method builds offline saves valid trajectories of lifting motion of the whole body, depending on the range of inertia parameters of a box, Ã¢ â¬ Han said. “Later, we developed an algorithm based on physical interaction that helps the robot interact with the box safely and estimate the inertia parameters of the box.”