Team Gets Robots to Reason If They Can Lift a Box Safely or Not

Team Gets Robots to Reason If They Can Lift a Box Safely or Not

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

The future seems bright for humanoid robots, there are a plethora of ongoing projects to get them to help humans with different tasks. Many of these projects focus on getting robots to pick up and carry objects for us.

There are many robots out there that can pick up and haul light objects of different shapes and sizes. But getting them to haul heavier objects has proved a hefty challenge. There's a possibility that the robot might drop or damage the object, or even hurt their back.


Ok, we might have made up the last part, but to address the other issue, a team of researchers at the Johns Hopkins University and National University of Singapore (NUS) came up with a process executed in three steps to help robots determine whether they'll be able to lift a given object or not.

Yuanfeng Han from the team told TechXplore "We were particularly interested in how a humanoid robot can reason about the feasibility of lifting a box with unknown physical parameters," and added, "To achieve such a complex task, the robot usually needs to first identify the physical parameters of the box, then generate a whole-body motion trajectory that is safe and stable to lift up the box."

Computational limitations

But, one potential problem is with the computational limitations. Generating motion trajectories to calculate a path for the lifting action is often "mentally taxing" for the robot. Oftentimes, humanoid robots have a quite high range of motion (ROM) and it needs to constraint their ROM to lift stuff.

That's why robots are often unable to complete their motion if the box is too heavy for them or the box's center of gravity is in an unexpected place.

Han explains "Think about us humans, when we try to reason about whether we can lift up a heavy object, such as a dumbbell, we first interact with the dumbbell to get a certain feeling of the object. Then, based on our previous experience, we kind of know if it is too heavy for us to lift or not."

"Similarly, our method starts by constructing a trajectory table, which saves different valid lifting motions 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 its previous experience."

The technique is developed by Han and his colleague Ruixin Li and it's supervised by mechanical engineering department professor Gregory Chirikjian. The robots first interact with the box briefly to sense its inertia parameters, then, the robot checks the trajectory table provided to it and determines whether it can successfully execute lifting the box.

If it finds a valid trajectory, it deems lifting the box feasible and goes for it. If not, the robot simply refuses the task. Or as Han explains: "Essentially, the trajectory table that our method constructs offline saves the valid whole-body lifting motion trajectories according to a box's range of inertia parameters. Subsequently, we developed a physical-interaction-based algorithm that helps the robot interact with the box safely and estimate the inertia parameters of the box."

This way robots also save computational power as they don't have to compute their motions before trying to lift something they can't. The article is prepublished on ArXiv.

Watch the video: Week in the Life. Household Tips, Whats for Dinner u0026 Working at Home! (February 2023).