A research team from the University of Cambridge has developed a passive robotic hand that can grip objects with the right amount of pressure using minimal energy. The robot hand senses the objects to be picked up and only uses its flexible wrist to grip it.
Conventional robotic hands pick up and hold objects with their movable fingers. This is done by an electric motor or pneumatically. However, both require a lot of energy. The University of Cambridge robotic hand used in the academic paper “Predictive Learning of Error Recovery with a Sensorised Passivity-based Soft Anthropomorphic Hand” described and published in Advanced Intelligent Systems, operates largely passively. The fingers are not moved, only the wrist is controlled by an electric motor.
Passive hand with flexible wrist
The anthropomorphic hand, 3D printed from soft material, is equipped with tactile sensors. The robotic hand can feel what it is touching. “This type of hand has a certain elasticity: it can pick things up by itself without using the fingers,” says lead author Dr. Kieran Gilday, who now works at the École Polytechnique Fédérale de Lausanne (EPFL) in Lausanne, Switzerland. “The tactile sensors give the robot a sense of how well the grip is working, so it knows when it’s starting to slip. This allows it to predict when it’s going to fail.”
The robotic hand learned through trial and error what type of grip to use to pick up a specific object. After training with small printed plastic balls, the researchers had the hand grasp different objects: a peach, a computer mouse and a roll of bubble wrap. In the tests, the robotic hand was able to successfully grab 11 out of 14 objects. In total, the team ran 1,200 tests.
“The robot learns that a combination of a specific movement and a specific set of sensor data will lead to failure, making it an adaptable solution,” says Gilday. “The hand is very simple, but it can pick up many objects with the same strategy.”
High energy efficiency
The advantage of the passive robot hand lies in its large range of motion, which can be achieved without the use of actuators. The hand is thus kept as simple as possible and yet it is able to show a complex behavior with high energy efficiency and low control effort, say the scientists.
The researchers are already thinking ahead. For example, the robotic hand could use computer vision and its environment to grasp an even wider range of objects.
(olb)
