Helping Robots Get a Grip
|Source: University of Utah's Robotic Grasp System- Precision (left), Power (right)|
Robot Grasping Planner System to Extend Robot Grasping Abilities
Scientists at the University of Utah have just developed a new, supervised learning system to enable robots to grasp objects with delicate precision and/or power/strength. The purpose is to greatly expand their functionality. It's called a probabilistic grasp planner and it's new innovation for the development of robotics.
Real Time Robot Planning Device
The device is able to plan grasps in real-time. The system models and plans high quality precision and strength grasps. It allows the robot to learn the best grasping technique as it practices on different objects. So for instance, the power and strength needed to pick up and place heavy objects and the precision and delicate manipulation needed for smaller, fragile objects.
Learning and Planning Grasp System for Robots
The collected data is used to determine which grasp is the most effective for each object. It also enables the robot to choose a grasp for unfamiliar objects. In essence, it gives robots a more precise and delicate sense of touch for the objects it is picking up, working with and delivering. For more news stories on innovation, go to amazon.com/author/ekane