How hands grasp and feel, and how robots learn to do the same.

What makes the human hand special

Pick up a coffee cup, button a shirt, turn a key: the hand does all of this without conscious effort. Behind it sit 27 bones, more than 30 muscles and tendons that run from the forearm into the fingers, and a skin that is one of the most sensitive surfaces of the body. The fingertips alone carry a very high density of mechanoreceptors that respond to pressure, stretch and vibration.

What makes the hand powerful is not any single part but the combination of a compliant structure, a smart nervous system and constant feedback. Skin and soft tissue deform around an object, receptors report where and how hard the contact is, and the brain adjusts the grip within a fraction of a second. Researchers sometimes describe this as an embodied solution: part of the intelligence sits in the body itself, not only in the brain.

Why robots struggle to copy it

Early robot grippers were rigid jaws with one motor. Later research hands tried to copy the anatomy joint by joint, with a motor for nearly every degree of freedom. These hands were impressive in the lab but heavy, costly and hard to control, because a computer had to calculate the position of every finger for every object.

Since about 2010 a different approach has gained ground. Instead of controlling everything, designers let the mechanics do part of the work: elastic joints, tendons that close several fingers at once and soft materials that wrap around an object. Our explainer on soft and underactuated robotic hands shows how this works, and the page on hand synergies explains the neuroscience idea behind it.

Touch: the missing half of grasping

Moving the fingers is only half the job. To hold an egg or a glass, you need to know how hard you press and whether the object starts to slip. People with numb fingertips drop things more often and grip harder than necessary. Robots face the same problem, which is why tactile sensors and haptic feedback are an active field. The explainer on haptic perception and touch feedback covers how touch works, how tactile sensors measure it and how haptic devices pass it on to a human operator.

From the lab to the arm: bionic hands

The most personal use of all this research is the prosthetic hand. Modern myoelectric hands read muscle signals from the residual limb, and some designs borrow the adaptive, soft grasping principles developed for robots. Our guide to bionic and myoelectric prosthetic hands explains the main types, how they are controlled, what daily life with them is like and which questions to discuss with a prosthetist. It does not replace medical advice.

How to read this site

Each explainer starts with the key points, then goes into the detail. Technical terms are explained where they first appear. We describe research results and products as they are publicly documented and avoid hype: many systems that work in a video are still far from daily use.